Technical Field
[0001] The invention relates to planning delivery of radiation. Particular examples provide
methods and systems for planning dose distributions for radiation therapy.
Background
[0002] Radiation is used in the treatment of cancer as well as some other medical conditions.
When radiation interacts with tissue, energy from the radiating particles is transferred
and deposited within the tissue. The energy is normally deposited in the vicinity
of the transfer. The maximum deposition is normally close to the point of interaction.
The energy deposited causes damage to cells that may eventually lead to cell death.
The quantity of energy deposited is normally described as radiation dose and has the
units of Gray (Gy). 1 Gray is equal to 1 Joule per kilogram of medium. The primary
goal of radiation treatment is to eradicate cancerous cells in a subject by depositing
sufficient radiation dose.
[0003] Radiation dose can damage or kill both cancerous and healthy tissue cells. It is
typical that some healthy tissue will receive radiation dose during a radiation treatment.
For example, a radiation beam originating from a radiation source and projecting through
a subject will deposit radiation dose along its path. Any healthy tissue located within
the path will normally receive some radiation dose. Additionally, some radiation dose
will typically be deposited outside of the beam path into healthy tissue due to radiation
scatter and other radiation transport mechanisms. One of the challenges of radiation
therapy is to deposit dose in cancerous tissue while minimizing dose received by healthy
tissue. Furthermore, some healthy tissues are more sensitive to radiation dose than
others making it more important to avoid radiation dose in those tissues.
[0004] Modern radiation delivery systems are capable of delivering complex dose distributions.
There is a desire for the radiation therapy clinician to be capable of evaluating,
determining and/or optimizing trade-offs between delivering dose to a tumor and minimizing
dose delivered to healthy tissue. Current techniques for evaluating these trade-offs
(treatment plan optimization, for example) are cumbersome and disconnect the operator
from quick and direct manipulation and evaluation of achievable dose distributions.
There is a desire for improvement of systems and methods for estimating achievable
dose distributions and possibly improving the evaluation of trade-offs between radiation
dose to cancerous and healthy tissue.
Summary
[0005] The following embodiments and aspects thereof are described and illustrated in conjunction
with systems, tools and methods which are meant to be exemplary and illustrative,
not limiting in scope. In various embodiments, one or more of the above-described
problems have been reduced or eliminated, while other embodiments are directed to
other improvements. The present invention provides a computer implemented method for
permitting manipulation of an achievable dose distribution estimate deliverable by
a radiation delivery apparatus for proposed treatment of a subject and a data processing
system for the same, as claimed.
[0006] In addition to the exemplary aspects and embodiments described above, further aspects
and embodiments will become apparent by reference to the drawings and by study of
the following detailed descriptions.
Brief Description of Drawings
[0007] Exemplary embodiments are illustrated in referenced figures of the drawings. It is
intended that the embodiments and figures disclosed herein are to be considered illustrative
rather than restrictive. In drawings which illustrate non-limiting examples of the
disclosure:
Figure 1 schematically depicts an example radiation delivery apparatus that may be
used in delivering radiation dose to a subject;
Figure 2 schematically depicts another example radiation delivery apparatus that may
be used in delivering radiation dose to a subject;
Figure 3 is a flowchart depicting a method for treatment planning comprising an optimization
process which is suitable for use with intensity modulated radiation treatment according
to a particular example;
Figure 4A is a flowchart depicting a method for planning radiation treatment and treating
a subject using radiation therapy involving generating, and permitting operator manipulation
of, estimated dose distribution according to an embodiment of the invention.
Figure 4B is a flowchart depicting a method for generating, and permitting operator
manipulation of, an estimated dose distribution according to a particular embodiment.
Figure 4C is a flowchart depicting an initialization method suitable for use in the
method of Figure 4B.
Figure 4D is a flowchart depicting a dose-estimation update method for determining
the changes in estimated dose distribution which may be used with the method of Figure
4B according to a particular example.
Figure 5A shows a schematic depiction of a two-dimensional cross section of a calculation
grid superimposed on image data which includes exemplary healthy tissue and target
structures together with the estimated dose for each voxel in the calculation grid.
Figure 5B shows a dose volume histogram (DVH) corresponding to the Figure 5A dose
distribution.
Figure 6 is an illustration of a beam between a radiation source and subject including
associated intensity grid and example ray lines.
Figure 7A shows an example of an operator communicating a desired dose change directly
on a representation (e.g. a graphical representation) of a dose distribution. Figure
7B shows a cross-section of the updated dose distribution determined in accordance
with the method of Figure 4B as a result of the Figure 7A desired dose change.
Figure 8A shows an example of an operator communicating a desired dose quality metric
change on a representation (e.g. a graphical representation) of a DVH. Figure 8B shows
the updated DVH determined in accordance with the method of Figure 4B as a result
of the Figure 8A desired dose change.
Figure 9A shows the dose modification voxels determined in accordance with the method
of Figure 4B and corresponding to the Figure 8A desired dose quality metric change.
Figure 9B shows a cross-section of the updated dose distribution determined in accordance
with the method of Figure 4B as a result of the Figure 9A desired dose change.
Figure 10 shows an example of an operator communicating a desired dose quality metric
change on a representation (e.g. a graphical representation) of a biological index.
Figure 11A shows the dose modification voxels determined in accordance with the method
of Figure 4B and corresponding to the Figure 10 desired dose quality metric change.
Figure 11B shows a cross-section of the updated dose distribution determined in accordance
with the method of Figure 4B as a result of the Figure 11A desired dose change.
Figure 12 schematically depicts an example of a dose update grid illustrative of a
technique which may be used to update the dose distribution as part of the method
of Figure 4B according to a particular example.
Figure 13 is a schematic illustration of a method for rapid estimation of achievable
dose distribution according to a particular example.
Figure 14 schematically depicts a 360° trajectory, a plurality of beams at angularly
evenly spaced locations around the trajectory, ray lines intersecting a centrally
located voxel and a two-dimensional cross-section of a dose update grid estimated
to be achievable by the plurality of beams.
Figure 15 depicts an example profile of the Figure 14 achievable dose update grid
intersects the dose modification voxel.
Figure 16 depicts a dose update grid which is the same as the Figure 14 dose update
grid except that it has been translated to a different dose modification voxel.
Figure 17 is a schematic depiction of a system for estimation and manipulation of
estimated dose according to a particular example of the invention.
Figures 18A-18C schematically depict a number of radial modification dose (RMD) distributions
and how various angular ranges of RMDs can be obtained by combining other angular
ranges of RMDs.
Figure 19A is a three-dimensional rendering showing a target structure and a healthy
tissue structure. Figure 19B is a cross-sectional view bisecting (and showing outlines
of) the target structure and healthy tissue structure of Figure 19A.
Figure 20A shows a representation of a cross-section of an initial estimated dose
distribution for the Figure 19A tissue structures. Figure 20B shows an initial DVH
for the Figure 19A tissue structures.
Figure 21A shows a magnified portion of the Figure 20B target structure DVH and operator
manipulation thereof. Figure 21B shows a number of dose modification distributions
resulting from the Figure 21A operator manipulations. Figure 21C shows an updated
dose distribution and updated target and healthy structure DVHs after the Figure 21A
manipulation.
Figure 22A shows operator manipulation of the Figure 20B healthy tissue DVH. Figure
22B shows a number of dose modification distributions resulting from the Figure 22A
operator manipulations. Figure 22C shows updated target and healthy structure DVHs
after the Figure 22A manipulation.
Figure 23A shows two example dose modification distributions used to correct the Figure
22C violation of the constraint on the target dose distribution. Figure 23B shows
target structure DVH after application of the Figure 23A dose modification distributions.
Figure 23C shows a representation of the resultant cross-sectional dose distribution.
Figure 24A shows DVHs for target and healthy tissue structures at the conclusion of
a series of dose modifications by the operator. Figure 24B shows a cross-sectional
estimated dose distribution corresponding to the DVHs of Figure 24A.
Description
[0008] Throughout the following description specific details are set forth in order to provide
a more thorough understanding to persons skilled in the art. However, well known elements
may not have been shown or described in detail to avoid unnecessarily obscuring the
disclosure. Accordingly, the description and drawings are to be regarded in an illustrative,
rather than a restrictive, sense.
[0009] Figure 1 schematically depicts an example radiation delivery apparatus 10 containing
a radiation source 111 for delivery of radiation to a subject 110 (e.g. a cancer patient).
In some examples, radiation source 111 may comprise, or may be generated by, a linear
accelerator. In general, radiation source 111 may be generated using any suitable
technique. For example, radiation source 111 may comprise decaying matter (e.g. Cobalt-60
or the like) or other types of source(s) which may emit radiation comprising neutrons,
electrons, protons, other charged particles and/or the like.
[0010] In the exemplary Figure 1 radiation delivery apparatus 10, radiation source 111 is
mounted to a gantry 112 and subject 110 is placed on table 113. Gantry-mounted radiation
source 111 and table 113 with subject 110 may rotate and/or translate with respect
to each other. For example, gantry 112 may rotate about a longitudinal axis 114 and
table 113 holding subject 110 may rotate about a vertical axis 115. Table 113 may
also translate relative to gantry 112 and source 111 in one or more of the three dimensions
shown by axes 116. Radiation from a radiation source may be emitted in all directions.
An enclosure 117 surrounds most of radiation source 111, so that the majority of radiation
emitted by source 111 is absorbed within enclosure 117.
[0011] Radiation emitted toward subject 110 is permitted to pass through enclosure 117 and
through a beam-shaping system 118. Beam-shaping system 118 may comprise one or more
collimators which may be used to define a beam of radiation that emanates from radiation
source 111 and into subject 110. The collimators of beam-shaping system 118 may be
motorized and their position and/or movement may be controlled (e.g. by a suitably
configured computer control system or the like). The collimators of beam-shaping system
118 may be controllably configured so that the shape of the radiation beam entering
subject 110 preferentially intersects cancerous tissue. In the illustrated example
radiation delivery apparatus 10 of Figure 1, beam-shaping system 118 comprises a multi-leaf
collimator (MLC) 118A having a plurality of individually controllable leaves. MLC
118A may also be controllably moved to rotate about beam axis 119.
[0012] The Figure 1 exemplary radiation delivery apparatus 10 can be used to deliver radiation
therapy treatment to subject 110. A number of techniques are known for using radiation
delivery apparatus 10 in manners which attempt to provide desired dose to diseased
tissue (e.g. cancerous tumor) while attempting to minimize the dose to healthy tissue.
A first such technique involves moving gantry 112 and/or table 113 to a configuration
wherein the radiation beam emitted from source 117 preferentially passes through relatively
more diseased tissue and relatively less healthy tissue. This first technique may
be improved by treating subject 110 with a first beam at a first configuration of
gantry 112 and table 113, moving the gantry 112 and table 113 relative to one another,
and then treating subject 110 with a second beam which projects from a different direction
than the first beam. Using this two-beam technique, the diseased tissue (tumor) may
receive a dose that is a combination of the dose from the two beams, while much of
the healthy tissue surrounding the tumor would receive dose from primarily a single
one of the beams. This two-beam technique may be extended to multiple (e.g. more than
two) beams. Each different relative orientation of gantry 112 and table 113 may correspond
to a different beam direction and may be referred to simply as a "beam".
[0013] Another example of a technique for using radiation delivery apparatus 10 in a manner
which attempts to provide desired dose to a tumor of diseased tissue while minimizing
dose to healthy tissue involves collimating each radiation beam (i.e. the beam from
each relative configuration of gantry 112 and table 113) so that the projection of
the tumor from the view of the radiation source (Beam's eye view) closely approximates
the outline of the tumor. In this way dose to healthy tissue surrounding the tumor
will be reduced. Collimation system 118 (e.g. MLC 118A) may be used to collimate the
individual beams. This collimation technique may be improved by selectively (partially
or fully) blocking portions of a radiation beam (from a first direction) that intersect
both tumor and sensitive healthy tissue and then compensate for the blocked portion
of the tumor by selectively (partially or fully) unblocking portions of one or more
radiation beams from one or more other directions. A radiation beam (from a particular
direction) comprising spatially non-uniform transmitted portions may be referred to
as "Intensity Modulated" in reference to the spatially varying intensity of radiation
across the two-dimensional beam projection. Intensity modulation can further improve
the difference between dose received by healthy tissue and dose received by tumor,
particularly in circumstances where some healthy tissue is of relatively high importance
(e.g. healthy organs) and/or is in relatively close proximity to the target tissue
and it is desired to impart even less dose to such highly important tissues.
[0014] In accordance with other techniques, it may be beneficial to dynamically move one
or more components of radiation delivery apparatus 10 during the delivery of radiation.
For example, collimation system 118 can change the shape of a radiation beam while
source 117 is emitting radiation, thereby providing dynamically varying collimation
shapes for intensity modulation. Additionally or alternatively, gantry 112 and table
113 can move relative to one another while source 117 is emitting radiation, thereby
providing continuously varying beam directions (in contrast to a finite number of
discrete beam directions). Some techniques may involve dynamically varying the position
of radiation source 117 through a motion trajectory (e.g. relative movement of gantry
112 and table 113) while simultaneously dynamically varying the collimated beam shape
and/or the intensity of radiation source 117.
[0015] Radiation delivery apparatus 10 represents only one example of a radiation delivery
apparatus. Other types of radiation delivery apparatus may be used to deliver therapeutic
radiation to a subject. A number of non-limiting examples of radiation delivery apparatus
include CyberKnife™ (Accuracy Incorporated), Tomotherapy™ (Tomotherapy Incorporated)
and Gammaknife™ (Elekta AB). Figure 2 schematically depicts an example of a Tomotherapy™
radiation delivery apparatus 12. In apparatus 12, radiation source 121 moves in a
toroid-shaped gantry 122 about subject 120 who is laying on table 123. Subject 120
and table 123 may be translated relative to gantry 122 (and/or gantry 122 may be translated
relative to subject 120 and table 123) in one or more of the directions indicated
by axes 124. For example, table 123 may be moved into and out of gantry 122 while
radiation source 121moves in circumferential direction(s) 125 within gantry 122. Radiation
emitted by source 121 may project through collimation system 126 (e.g. a MLC 126A
or the like). Collimation system 126 may move circumferentially (with source 121)
within gantry 122 to create a fan beam 127 impinging on subject 120. As source 121
and collimation system 126 move in circumferential directions 125 around subject 120,
table 123 and subject 120 may move relative to gantry 120 and MLC 126A may be configured
to provide intensity modulated beams from varying directions around subject 120.
[0016] Using radiation delivery apparatus (like exemplary apparatus 10, 12 of Figures 1,
2) to deliver therapeutic radiation treatment involves treatment planning. Treatment
planning may involve determining the information used to control the radiation delivery
apparatus during radiation delivery. Such information may include, by way of non-limiting
example, beam configuration parameters (e.g. a number of beams, directions of beams,
radiation energy and/or the like) and beam delivery parameters (e.g. collimation shape(s)
and corresponding collimation system configuration(s), radiation intensity(s) and/or
the like).
[0017] Since the size, shape and position of a tumor with respect to the surrounding healthy
tissue are different for each subject, a diagnostic imaging procedure is typically
used prior to (or as a part of) treatment planning for the purposes of determining
the spatial locations of diseased and healthy tissue. Computed Tomography (CT), Magnetic
Resonance Imaging (MRI) and Positron Emission Tomography (PET) are common imaging
methods used for this diagnostic imaging process. The result of CT, MRI and PET imaging
may comprise 3-dimensional images which contain anatomical and functional information.
In some examples, diagnostic imaging may involve procuring 4-dimensional images, which
incorporate time varying image information (e.g. to account for breathing).
[0018] The locations of diseased and healthy tissue may be identified on these images prior
to (or as a part of) treatment planning. Identification of the locations of diseased
and healthy tissue can be performed manually although methods for automatic and semi-automatic
identification may also be used. Treatment planning may involve using well known methods
to model the radiation dose resulting from a radiation beam. The dose that a subject
would receive from a particular radiation beam may be evaluated by overlaying the
dose distribution modeled for the particular radiation beam on the subject's images.
The dose that a subject would receive from a proposed treatment plan (e.g. a plurality
of radiation beams) may be evaluated by superposing dose distributions modeled for
the individual beams of the plan and overlaying the superposed dose distribution on
the subject's images.
[0019] A proposed treatment plan may be characterized or otherwise specified by a set of
radiation delivery parameters. As used in this specification and the accompanying
claims, radiation delivery parameters may comprise: beam configuration parameters
which may relate to the geometric positioning of the radiation source with respect
to the subject (e.g. numbers of beams, directions of beams, radiation energy of beams,
motion of beams (e.g. for continuously varying beams) and/or the like) and/or beam
delivery parameters which may relate to the characteristics of one or more beam configurations
(e.g. collimation shape(s) and corresponding collimation system configuration(s),
motion of collimation shape(s) and corresponding collimation system configuration(s)
(e.g. for continuously varying collimation shapes), radiation intensity(s), and/or
the like). By estimating/modeling the dose distributions from multiple different treatment
plan proposals (as characterized by multiple corresponding sets of radiation delivery
parameters), the multiple different treatment plan proposals can be compared against
one another. Once a proposed treatment plan is selected (e.g. because it is determined
to be superior to others or is otherwise determined to be satisfactory), the radiation
delivery parameters associated with the selected treatment plan may be transferred
to the radiation delivery apparatus for delivery of the selected treatment plan to
the subject.
[0020] Treatment planning for intensity modulated radiation therapy (IMRT) may be more complex
because of the permissible spatial variation of intensity distribution across a two-dimensional
cross section of each beam or each portion of a beam. Because of the spatial varying
intensity distributions of beams associated with IMRT, IMRT treatment planning typically
involves dividing each beam into a two-dimensional matrix of spatially varying intensity
portions which may be referred to as beamlets. Each beamlet may be effectively treated
as a separate beam element that follows a ray line from the source into the subject.
In a typical non-limiting IMRT plan, there may be 5 to 9 beams each with a matrix
that may comprise more than 100 beamlets.
[0021] IMRT planning by a human observer is normally considered impractical due to the large
number of beamlets. Several computer algorithms have been developed to determine the
spatially varying intensity distributions of each beam in IMRT plans - e.g. the intensities
for each of the beamlets in each portion of each beam of the IMRT plan. These algorithms
typically involve iterative optimization. For example, at each iteration, a particular
set of radiation delivery parameters is proposed, the corresponding dose distribution
is calculated (e.g. modeled) and the corresponding dose distribution is evaluated
by comparing a quality metric associated with the corresponding dose distribution
to some objective. The next iteration may then attempt to propose a set of radiation
delivery parameters whose dose distribution is superior to the previous iteration
(when evaluated in relation to the same objective). The iterative process normally
repeats until an optimization termination criterion is achieved. The iterative optimization
process may be relatively time consuming, computationally expensive and/or temporally
inefficient, because of the need to model/calculate a dose distribution at each iteration.
[0022] Figure 3 schematically depicts a method 14 for treatment planning suitable for use
with IMRT according to a particular example. Method 14 commences in block 130 which
involves delineating target (e.g. diseased) tissue and healthy tissue within the subject
using whatever diagnostic image information may be available. In some examples, block
130 may involve sub-dividing healthy tissue into one or more levels of relatively
important healthy tissue (e.g. organs) and relatively less important healthy tissue.
In some examples, block 130 may involve acquiring image information from (or through
effecting) an imaging procedure. Such image information may be obtained by any suitable
technique, including (without limitation): Computed Tomography (CT), Magnetic Resonance
Imaging (MRI), Positron Emission Tomography, Ultrasound and/or any other suitable
imaging procedure. The image information may be used in block 130 to determine spatial
anatomical and functional information regarding diseased and healthy tissue in the
subject. This delineation between target and healthy tissues and structure identification
may be performed manually, by an operator, and/or using automatic or semi-automatic
image segmentation methods.
[0023] Block 131 involves specifying dose objectives for target and healthy tissues. Ideally,
the dose objectives might be: (i) 100% of the tumor volume receives at least the prescription
dose; (ii) 0% of the tumor volume receives dose greater than the prescription dose;
and (iii) 0% of the healthy tissue volume receive any dose. Such an ideal objective
is not realizable in practice. Instead, achieving a prescription dose to the tumor
must be balanced against providing a low dose to healthy tissue structures. Also,
there are commonly many healthy tissue structures of concern that vary in importance
which can make the number of possible trade-offs cumbersome. By way of non-limiting
example: a healthy tissue objective may comprise maximum 20% of the healthy tissue
structure volume receives 30% of the dose prescribed to the target; and a target tissue
objective may comprise minimum of 90% of the tumor target volume receives 95% of the
dose prescribed to the target. Block 131 may also involve specifying a dose quality
metric which may be used to evaluate proposed treatment plan iterations against the
block 131 dose distribution objectives. Such dose quality metrics may additionally
or alternatively be used to indirectly specify dose distribution objectives.
[0024] Method 14 then proceeds to block 132 which involves performing an iterative optimization
process to arrive at a set of radiation delivery parameters. A typical iterative optimization
process which may be implemented as a part of block 132 was discussed above. At the
conclusion of the block 132 iterative optimization process, method 14 proceeds to
block 133 which involves evaluating the treatment plan (e.g. the calculated/modeled
dose distribution) which results from using the radiation delivery parameters output
by the block 132 optimization. The block 133 evaluation may be performed by one or
more human operators. In some examples, however, the block 133 evaluation may be automated.
For example, constraints could be developed that specify a minimum number of objectives
that must be achieved in order to achieve a positive block 133 evaluation. The number
of achieved objectives might be more important for some target structures or for some
specified healthy tissue structures. Even in the case of an automated block 133 evaluation,
the resulting dose distribution would likely be evaluated by a clinician before actually
delivering radiation.
[0025] If the block 131 dose distribution objectives are not achievable in practice, then
the block 132 optimization may fail to determine an acceptable treatment plan. Conversely,
if the block 131 dose distribution objectives are too easily achieved, the treatment
plan specified by the block 132 optimization may not achieve the best trade-offs that
could be realized in practice. In these circumstances (block 133 NO output path),
method 14 may loop back to block 131, where a different optimization may be performed
with different dose distribution objectives. Typically, to ensure that the dose distribution
achieved in treatment planning method 14 is close to optimal, it is desirable to perform
multiple block 132 optimizations with multiple sets of block 131 dose distribution
objectives. Once it is determined in block 133 that a particular plan is optimal (block
133 YES output path), then method 14 may proceed to block 134 where the radiation
delivery parameters may be transferred to the radiation delivery apparatus for delivery
to the subject.
[0026] Other methods of providing complex treatment plans and corresponding dose distributions
may involve similar iterative optimization. Any combination of radiation delivery
parameters may be used in an optimization process for deriving a treatment plan. Such
complex treatment plans and corresponding dose distributions may include, by way of
non-limiting example, so-called direct aperture optimization techniques, radiation
treatment techniques involving dynamic variation of the direction of radiation (e.g.
movement of the radiation source during treatment) and/or dynamic variation of collimation
shapes and/or intensities within particular beams during treatment and radiation treatment
techniques using radiation delivery apparatus such as CyberKnife™ (Accuracy Incorporated),
Tomotherapy™ (Tomotherapy Incorporated) and Gammaknife™ (Elekta AB). The iterative
optimizations involved in treatment planning for all of these radiation delivery techniques
suffer from similar drawbacks as those discussed above for the iterative optimization
associated with IMRT treatment planning. More particularly, such optimizations are
time consuming and computationally expensive because of the need for calculating/modeling
a dose distribution at each iteration.
[0027] One aspect of the invention provides systems and methods for permitting manipulation
of achievable dose distribution estimates. In particular embodiments, estimated dose
distributions and associated dose quality metrics may be manipulated without the cumbersome
and computationally expensive calculations involved in simulating dose for specific
radiation delivery parameters (e.g. without the need for iterative optimization).
These methods and systems may be simple to use and may permit operator manipulation
of estimated dose distributions and associated dose quality metrics. By way of non-limiting
example, an operator may select a graphical representation of a dose quality metric
using a computer mouse or similar computer pointing device, drag it to the left or
right, up or down as desired. As another non-limiting example, an operator may modify
a graphical representation of a dose distribution using a computer mouse or similar
computer pointing device, to "paint" or "erase" dose from a region of subject anatomy.
As these operator-directed manipulations are made, the achievable dose distribution
estimate and corresponding dose quality metrics may be updated in near real-time.
[0028] The range of physically achievable dose distributions may be limited. Particular
embodiments involve the imposition of limits or restrictions on available manipulations,
so that the estimated dose distributions (after operator manipulation) are at least
approximately achievable. In this way operators are able to rapidly explore trade-offs
between dose delivery to target tissues (e.g. tumor(s)) and healthy tissues (e.g.
organ(s)) while ensuring that the subject will ultimately receive a dose distribution
substantially similar to the estimated one.
[0029] Figure 4A schematically depicts a method 16 for planning radiation treatment and
treating a subject using radiation therapy involving manipulation of estimated dose
distribution. Method 16 may generally be divided into two parts: a first part 16A
which involves planning the radiation treatment and a second part 145 which is not
part of the claimed invention and involves delivering radiation treatment to the subject.
As discussed further herein, planning part 16A may involve determining radiation delivery
parameters which may be provided to a radiation delivery apparatus to permit block
145 delivery of radiation in accordance with the plan.
[0030] Method 16 commences in block 141 which involves delineating target (e.g. diseased)
tissue and healthy tissue within the subject using whatever diagnostic image information
may be available. Block 141 may be substantially similar to block 130 described above
for method 14 (Figure 3). Method 16 then proceeds to block 142 which involves generating
and permitting operator manipulation of an achievable dose distribution. Operator
manipulation of the achievable dose distribution permitted as a part of block 142
may be analyzed directly, effectively in real-time, facilitating a rapid and more
comprehensive understanding of the compromises between target tissue dose and healthy
tissue dose. Operator manipulation of the achievable dose distribution permitted as
a part of block 142 may comprise modification of or additions to target tissue and/or
healthy tissue. Block 142 of method 16 is described in more detail below.
[0031] At the conclusion of block 142 (e.g. where the operator is satisfied with the manipulated
version of the achievable dose distribution or otherwise), method 16 proceeds to optional
block 143 which involves determining radiation delivery parameters capable of permitting
a radiation delivery apparatus to deliver the estimate of achievable dose as output
from block 142. Block 143 may involve performing an iterative optimization process
or the like to derive radiation delivery parameters (e.g. beam configuration parameters
and/or beam delivery parameters). The block 143 iterative optimization may involve
processes similar to those described in blocks 131 and 132 of treatment planning process
14 described above (Figure 3). The block 143 optimization may also involve an evaluation
similar to that of block 133 of treatment planning process 14 described above (Figure
3). Advantageously, because of the availability of dose distribution manipulation
in block 142, it may not be necessary to perform multiple optimizations with different
dose distribution objectives as a part of block 143 or fewer optimizations with different
dose distributions may be performed as a part of block 143 when compared to the optimization
process (blocks 131, 132, 133) of method 14 (Figure 3) - i.e. the availability of
dose distribution manipulation in block 142 of method 16 may reduce or eliminate the
need for multiple iterative optimization loops (analogous to loops through blocks
131, 132 and 133 (NO output path) of the method 14 optimization process).
[0032] After optimization to obtain the radiation delivery parameters in block 143, method
16 proceeds to block 144 which involves transferring the block 143 radiation delivery
parameters to the controller of a radiation delivery apparatus. These radiation delivery
parameters may then be used by the controller of the radiation delivery apparatus
in block 145 to cause the radiation delivery apparatus to deliver radiation to the
subject in accordance with the radiation treatment plan corresponding to the radiation
delivery parameters.
[0033] As discussed briefly above, block 142 of radiation treatment method 16 involves generating,
and permitting operator manipulation of, an achievable dose distribution. Figure 4B
schematically illustrates a method 18 for generating, and permitting operator manipulation
of, an achievable dose distribution according to a particular embodiment. Method 18
of Figure 4B may be used to implement block 142 of radiation treatment method 16 (Figure
4A).
[0034] Method 18 commences with initialization in block 218. The block 218 initialization
may involve: establishing a calculation grid over a region of interest in the delineated
image information; defining a configuration of beams; defining an initial intensity
distribution of beamlets for each beam; generating an initial estimate of an achievable
dose distribution using the beam configuration and beamlet intensity distributions;
and, optionally, determining an initial estimate of one or more dose quality metrics
based on the initial estimated dose distribution.
[0035] Figure 4C schematically illustrates an initialization method 20 suitable for use
in block 218 of method 18 according to a particular example. Initialization method
20 commences in block 40 which involves creating a calculation grid and superimposing
the calculation grid over the delineated image data and segmented healthy and target
tissue structures. The delineated image data may be obtained as a part of block 141
(Figure 4A) discussed above. The superimposed calculation grid may comprise a three-dimensional
grid of voxels that spans a region of interest within a subject. The three-dimensional
grid of voxels may be characterized by a suitable coordinate system which may permit
indexing and/or identifying individual voxels within the grid by their corresponding
coordinates.
[0036] Figure 5A shows a schematic depiction of a two-dimensional cross sectional portion
151 of a three-dimensional calculation grid 151A superimposed on image data comprising
exemplary anatomical structures including healthy tissue structure 152 (shown in dashed
lines) and target structure 153 (shown in full lines). Only two anatomical tissue
structures are shown in the Figure 5A example. In practice, there may be different
numbers of structures (target structures and/or healthy tissue structures) associated
with a region of interest in a subject. While the illustrated portion 151 of calculation
grid 151A may be referred to as a two-dimensional cross-section, each box of grid
portion 151 shown in Figure 5A actually represents a corresponding voxel in the overall
three-dimensional calculation grid 151A for the subject. In this sense, the Figure
5A grid portion 151 is actually three-dimensional grid portion 151 with a depth of
one voxel. As will be explained in more detail below, the numerical values in each
voxel of grid portion 151 represent example values of the dose distribution expected
to be delivered to these voxels, although these dose distribution estimates will not
typically be known when calculation grid is established in block 40.
[0037] Returning to Figure 4C, initialization method 20 proceeds to block 42 which involves
defining a configuration of beams. The block 42 initial configuration of beams may
comprise defining parameters similar to those beam configuration parameters discussed
above - e.g. numbers of beams, directions of beams, radiation energy of beams and/or
the like. The block 42 beam configuration may be identified manually or through an
independent optimization process. The block 42 beam configuration may comprise one
or more static beams, one or more continuously moving beams or a combination of static
and moving beams. Continuously moving beams may be characterized by a motion path
(trajectory) of the beam direction together with a suitable sampling of multiple stationary
beams with positions along the motion path.
[0038] Initialization method 20 then proceeds to block 44 which involves defining an initial
intensity distribution of beamlets for each of the block 42 beams. Such initial intensity
distributions may be similar to initial values for parameters similar to beam delivery
parameters discussed above. Figure 6 schematically depicts a representative beam 159
from the block 42 beam configuration. Beam 159 is directed from a radiation source
161 and delivered toward a subject 160. Each beam in the block 42 beam configuration
may be characterized, at least in part, by the position of radiation source 61 with
respect to subject 60 (as defined by the block 40 calculation grid and the corresponding
image data for subject 60) . Each beam 159 is associated with a corresponding intensity
distribution 165 which is defined in block 44. In block 44, the intensity distribution
165 corresponding to each beam 159 may be segmented into a two-dimensional grid 162
of intensity beamlets 164. Two dimensional intensity distribution grid 162 may be
characterized by a suitable coordinate system which may permit indexing and/or identifying
individual beamlets 164 within grid 162 by their corresponding coordinates. Each beamlet
164 may be associated with a corresponding ray line 163 which originates from radiation
source 161 and passes through intensity distribution 165 and grid 162 at the location
of the beamlet 164. In the Figure 6 illustration, a particular ray line 163A is shown
to correspond with a particular beamlet 164A.
[0039] Block 44 also involves assigning initial intensity values to the individual beamlets
164 for each beam 159 - i.e. initializing the intensity distribution 165 for each
beam 159. The initial intensity distributions 165 may be defined in block 44 using
a variety of different techniques. By way of non-limiting example:
- Intensity distribution 165 is zero for all beamlets 164 and all beams 159. This initial
intensity assignment corresponding to zero dose for all structures.
- Intensity distribution 165 is random over each grid 162 of beamlets 164 for each beam
159, intensity of each beamlet 164 is random over all intensity distributions 165
for all beams 159 or some other suitable scheme involving at least some form of random
assignment of intensities to corresponding beamlets 162.
- Intensity distributions 165 for all or a subset of beams 159 are assigned so that
an approximate sphere of dose encompasses target structure(s) 153.
- Intensity distributions 165 for all or a subset of beams 159 are assigned such that
beamlets 162 corresponding to ray lines 163 that intersect target structure(s) 153
are assigned an initial positive intensity value (e.g. unity) and beamlets 162 are
otherwise assigned a different initial value (e.g. zero or some other relatively low
value).
- Intensity distributions 165 for all or a subset of beams 159 are rescaled (e.g. by
an equal amount) so that the resultant estimated dose distribution would have a maximum
dose corresponding to some dose threshold (e.g. a dose threshold equal to a highest
prescription dose for target structure(s) 153).
- Intensity distributions 165 for all or a subset of beams 159 are proportional or correlated
with a desired (e.g. prescription) dose for target structure(s) 153.
∘ Intensity distributions 165 are generated for all or a subset of beams 159 where
beamlets 162 corresponding to ray lines 163 that intersect target structure(s) 153
are assigned an intensity value proportional to or correlated with a desired dose
for each such target structure 153.
∘ A margin (e.g. offset) may be added to the intensity distributions 165 (or to individual
beamlets 162 corresponding to ray lines 163 that intersect target structure(s) 153)
to ensure proper coverage of the target structure(s) 153 by the resulting dose distribution.
∘ For target structures 153 that overlap in a given beam 159 (e.g. a ray line 163
intersects multiple target structures 153), the intensity of the corresponding beamlet
164 may be assigned an intensity proportional to or correlated with the desired dose
for the target structure with the highest desired dose and/or the corresponding beamlet
164 may have its assigned intensity weighted by the target structure 153 that has
the highest desired dose.
∘ For target structures that overlap in a given beam 159, the intensity of the corresponding
beamlet 164 may be assigned an intensity proportional to or correlated with the desired
dose for the target structure with the lowest desired dose and/or the corresponding
beamlet 164 may have its assigned intensity weighted by the target structure 153 that
has the lowest desired dose.
[0040] Returning to Figure 4C, after defining intensity distributions 165 in block 44, method
20 proceeds to block 46 which involves generating an initial estimate of the dose
distribution based on the block 42 beam configuration and the block 44 intensity distributions.
Block 46 may involve known methods of dose estimation, such as by way of non-limiting
example, Monte Carlo, collapsed cone convolution, pencil beam, anisotropic analytical
algorithm, Boltzman equation solvers and/or the like. The block 46 dose estimation
may comprise independently estimating the dose for each beam 159 of the block 42 beam
configuration and then adding these dose contributions to arrive at an overall initial
dose distribution estimate. The block 46 dose estimation may involve assigning a dose
value to each voxel in the block 40 calculation grid 151A. These dose estimate values
are represented in Figure 5A by the numbers in the boxes corresponding to voxels,
it being understood that higher numbers correspond generally to higher estimated dose
amounts. Initialization method 20 may optionally involve a procedure (in block 47)
for establishing an approximate relationship between the intensities of beamlet(s)
164 having ray line(s) 163 that intersect a particular voxel and the corresponding
dose delivered to the particular voxel. This block 47 approximate relationship may
be determined using one or more suitable calibration procedures, may be determined
based on empirical testing and/or data and/or the like. This block 47 relationship
is explained in more detail below.
[0041] Method 20 may then proceed to block 48 which involves optionally determining one
or more initial dose quality metrics based on the block 46 initial dose distribution
estimate. Dose quality metrics determined in block 48 may generally comprise any function
of the estimated dose distribution. Some dose quality metrics include:
- The average dose to a structure;
- Dose volume histogram(s) - often referred to as DVHs;
- Rate of dose fall-off outside target structure(s) 153;
- Dose conformity indices - e.g. how closely the prescription dose matches the shape
of target structure(s) 153);
- Radiobiological objective(s) -e.g. tumor control probability, normal tissue complication
probability, equivalent uniform dose, and/or the like.
[0042] DHVs represent one popular and widely used dose quality metric. A DVH is a graphical
plot of structure volume (target tissue or healthy tissue) on the Y-axis versus dose
on the X-axis. It is common to use cumulative DVHs (which are typically referred to
simply as 'DVHs') when evaluating treatment plans. Figure 5B shows a plot 154 of two
typical DVHs corresponding to the estimated dose distribution shown in the three-dimensional
calculation grid 151A for which a cross-sectional portion 151 is shown in Figure 5A.
DVH 156 corresponds to target structure 153 and DVH 155 corresponds to healthy tissue
structure 152.
[0043] Returning to the Figure 4B method 18 for generating and permitting manipulation of
achievable dose distributions, after completion of the block 218 initialization, method
18 proceeds to block 220 which involves determining the coordinates and estimated
magnitude of desired dose modifications. The coordinates determined in block 220 to
be associated with desired dose changes may be referred to as the desired dose modification
coordinate(s)/voxel(s) and the associated magnitudes may be referred to as the desired
dose modification magnitude(s). Block 218 may involve receiving operator input which
is indicative of desired dose modifications. By way of non-limiting example, an operator
may indicate a desired dose modification (increase or decrease) to a particular dose
modification voxel in calculation grid 151A, an operator may communicate a desired
modification to a dose quality metric and/or the like. Various examples may comprise
one or more of a variety of different techniques for receiving such operator input.
Such techniques may include (without limitation):
- Keyboard entry (e.g. typing a spatial location (e.g. calculation grid coordinates)
of the desired dose modification voxel and the amount of the desired dose modification
or typing a desired modification to a dose quality metric).
- Manipulation of graphical representations of dose distributions and/or dose quality
metrics via a graphical user interface.
- Specifying a location on a graphical representation of the subject using a mouse or
similar computer pointing device and, using a mouse button or keyboard input to indicate
whether to increase or decrease the dose at that location and/or to indicate the magnitude
of the desired dose modification.
- Other computer input devices (trackball, touch screen, voice command, video command
and/or the like) may also be used
[0044] Figure 7A shows an example of an operator communicating a desired dose modification
directly on a representation (e.g. a graphical representation) of a dose distribution
170. In the Figure 7A example, the operator selects a dose modification location (e.g.
dose modification voxel) 171 inside a healthy tissue structure 152 and then communicates
a desired dose reduction at that dose modification voxel171.
[0045] Figure 8A shows an example of an operator communicating a desired dose modification
on a representation (e.g. a graphical representation) of a DVH. In the Figure 8A example,
the operator selects a point 181 on the DVH 182 corresponding to a target structure
and indicates that an adjustment corresponding to a higher dose 183. As discussed
above, block 220 involves determining coordinates (e.g. voxel locations) and estimated
magnitudes of desired dose modifications. Accordingly, block 220 may involve a process
for converting the DVH input of Figure 8A to corresponding dose modification coordinates/voxels
and desired dose modification magnitudes. Similarly, block 220 may involve process(es)
for converting other input relating to other dose quality metric(s) to corresponding
dose modification coordinates/voxels and desired dose modification magnitudes. In
particular examples, the coordinates and magnitudes of desired dose modifications
may be determined in block 220 by processes which comprise:
- inverting the dose quality metric function, so that dose magnitude and dose coordinates
become function(s) of the dose quality metric; and
- calculating the required dose distribution modification (magnitude and location) by
applying the operator-indicated modifications to the dose quality metric value in
the inverted dose metric function.
[0046] Figure 9A shows part of the result of this block 220 inversion procedure which includes
the dose modification coordinates (e.g. voxel locations) 191 of the dose modifications
corresponding to the Figure 8A operator manipulation of the DVH. In the illustrative
example of Figure 9A, the Figure 8A DVH manipulation corresponds to desired increases
in the dose magnitude of coordinates (voxel locations) 191 inside target structure
153. Although the magnitudes of the corresponding dose modifications are also determined
as part of the block 220 inversion procedure, they are not explicitly shown in the
schematic depiction of Figure 9A.
[0047] Figure 10 shows an example of an operator communicating a desired dose modification
on a representation (e.g. a graphical representation) of another type of dose quality
metric (in the illustrated example, a biological index known as the Normal Tissue
Complication Probability (NTCP) and Tumor Control Probability (TCP)). In the Figure
10 example, the NTCP and TCP are displayed in the form of a bar histogram, the operator
selects NTCP 203 and indicates (at 201) that NTCP 203 should be reduced. As discussed
above, when the operator input is a desired modification in a dose quality metric,
block 220 may involve performing an inversion procedure to determine the coordinates
(e.g. voxel locations) and magnitudes of the desired dose modifications corresponding
to the operator manipulation. The results of this block 220 inversion are depicted
in Figure 11A which shows the coordinates (e.g. voxel locations) 211 of the dose modification
corresponding to the Figure 10 operator manipulation of the NTCP. In the illustrative
example of Figure 11A, the Figure 10 NTCP manipulation corresponds to desired decreases
in the dose magnitude of coordinates (voxel locations) 211 inside healthy tissue structure
152. Although the magnitudes of the corresponding dose modifications are also determined
as part of the block 220 inversion procedure, they are not explicitly shown in the
schematic depiction of Figure 11A.
[0048] Where desired dose modification input is received in the form of manipulation of
dose quality metrics, techniques other than inversion may be used to predict desired
dose modification voxels and corresponding dose modification magnitudes. In one particular
example, a change to a DVH is received as input which comprises a change to a point
on a DVH curve which may be identified by a corresponding dose (
D_selected) and a corresponding volume (
V_selected). The dose modification voxels may then be identified (in block 220) using a number
of techniques including: identifying voxels to be dose modification voxels if the
voxels have values falling within
D_selected +/-
Δ, where Δ may be a fixed value (which may be operator-selectable), a fraction of
D_selected (which may be an operator-selectable fraction), a value determined by calibration
or empirical evidence and/or the like. If no voxels are identified to be dose modification
voxels, then Δ may be expanded and voxels may be re-identified until at least one
dose modification voxel is identified. If a large number of voxels (e.g. all voxels
for that structure or a number of voxels greater than a threshold number or a threshold
percentage of the voxels in a structure) are identified to be dose modification voxels,
then Δ may be reduced and voxels may be re-identified. In some examples, all voxels
inside the structure for which the DVH is changed may be identified as dose modification
voxels. By way of non-limiting example, the magnitudes of dose modifications may be
determined on the basis of: a fraction of the
D_selected value; a fraction of the maximum dose for the structure to which the DVH corresponds;
a fraction of a prescription dose assigned to the structure to which the DVH corresponds;
a fraction of the maximum prescription dose assigned to all structures; a fraction
correlated with (e.g. proportional to) the amount of mouse or similar computer pointing
device movement by an operator; a operator-selected quantity; a fixed quantity which
may be an operator-configurable parameter or may be a "hard coded" constant; a combination
of the above; and/or the like.
[0049] Block 220 involves determining the coordinates (e.g. voxel location(s)) and magnitudes
of desired dose changes. The coordinates determined in block 220 to be associated
with desired dose changes may be referred to as the desired dose modification coordinate(s)/voxel(s)
and the associated magnitudes may be referred to as the desired dose modification
magnitude(s). Typically, although not necessarily, such desired dose modification
coordinate(s) and magnitude(s) are determined on the basis of operator input, but
could additionally or alternative be generated based on other forms of input (e.g.
computer-generated automated test input and/or the like). Such operator input may
involve direct specification of desired dose modification coordinates and magnitudes
or indirect specification of desired dose modification coordinates and magnitudes
through specification of desired changes to one or more dose quality metrics. The
examples shown in Figures 7A, 8A, 9A, 10A and 11A are for illustrative purposes only
and are not meant to limit the scope of the invention. There are many other dose quality
metrics that could be used. Furthermore, each dose quality metric could be represented
in many different formats (graphical or otherwise). The block 220 procedure may involve
the manipulation of any dose quality metric and representation that is a function
of the dose distribution.
[0050] In a specific embodiment, block 220 may involve determining secondary dose modification
coordinates/voxels and corresponding dose modification magnitudes in addition to the
primary dose modification coordinates/voxels and corresponding dose modification magnitudes
determined in accordance with the techniques described above. Secondary dose modification
voxels may be defined in a marginal region proximate to the primary dose modification
voxels. By way of non-limiting example, such secondary dose modification voxels could
be determined to be in a marginal region less than or equal to a threshold number
of voxels away from the primary dose modification coordinates. The threshold number
of voxels that define the marginal region may be operator-configurable, may be a system
parameter which may be determined by one or more suitable calibration procedures,
may be a system parameter which may be determined based on empirical testing and/or
data and/or the like.
[0051] The secondary dose modification magnitudes of the secondary dose modification voxels
may be less than the primary dose modification magnitudes of their corresponding primary
dose modification voxels. For example, the secondary dose modification magnitudes
may be a fraction
a of the primary dose modification magnitudes (where 0<=
a<=1). This fraction may be a function of the distance between a given secondary dose
modification voxel and its corresponding primary dose modification voxel -e.g. within
a marginal region of 3 voxels around a primary dose modification voxel, the fraction
a may be relatively high for the secondary dose modification voxels that are nearest
neighbors to the primary dose modification voxel; lower for the secondary dose modification
voxels that are spaced by one voxel from the primary dose modification voxel; and
lowest for the secondary dose modification voxels that are spaced by two voxels from
the primary dose modification voxel.
[0052] In other respects (e.g. for the purposes of other procedures involved in the methods
and systems described herein), secondary dose modification voxels determined as a
part of block 220 may be treated, for the most part, in the same manner as primary
dose modification voxels determined in block 220. Accordingly, both primary dose modification
voxels and secondary dose modification voxels determined in block 220 may be referred
to simply as dose modification voxels.
[0053] Returning to Figure 4B, method 18 then proceeds to block 221 which involves determining,
for each beam 159 in the block 42 beam configuration, the ray lines 163 that intersect
the voxels corresponding to the block 220 desired dose modification coordinates. As
discussed above, each of these ray lines 163 is associated with a corresponding beamlet
164 of a corresponding intensity distribution 165. Block 221 may also involve determining
the coordinates of these beamlets 164. The ray lines 163 that intersect the voxels
corresponding to the block 220 desired dose modification coordinates may be referred
to as dose-change ray lines 163 and their corresponding beamlets 164 may be referred
to as the dose-change beamlets 164.
[0054] Method 18 then proceeds to block 223, which involves adjusting the intensities of
the dose-change beamlets 164 identified in block 221. For example, if it is determined
in block 220 that the dose corresponding to a particular desired dose modification
voxel is to be decreased, then block 223 will typically involve decreasing intensities
of the corresponding dose-change beamlets 164. Conversely, if it is determined in
block 220 that the dose corresponding to a particular desired dose modification voxel
is to be increased, then block 223 will typically involve increasing intensities of
the corresponding dose-change beamlets 164.
[0055] Changes to the intensities of the dose-change beamlets 164 in block 223 may be effected
using a wide variety of techniques. By way of non-limiting example:
- identical magnitude intensity changes or identical percentage intensity changes may
be applied to each dose-change beamlet 164;
- identical intensity changes may be applied to one or more subsets of dose-change beamlets
164. Such subsets of dose-change beamlets 164 could correspond, for example, to consecutive
dose-change ray lines 163, to dose-change beamlets 164 from every nth (e.g. every second or third) beam 159, to randomly selected dose-change beamlets
164 and/or the like;
- different magnitude or percentage intensity changes may be applied for all or a subset
of dose-change beamlets 164; and/or
- the like.
[0056] The magnitudes of the block 223 changes to the intensities of the dose-change beamlets
164 is a function of (e.g. correlated with or proportional to) the block 220 desired
dose modification magnitude(s). As discussed briefly above, initialization method
20 may optionally include a procedure (block 47) which establishes an approximate
relationship between the intensities of beamlet(s) 164 having ray line(s) 163 that
intersect a particular voxel (e.g. dose-change beamlets 164 that intersect a dose
modification voxel) and the corresponding dose delivered to the particular voxel.
While this approximate relationship is shown as being determined in block 47 of the
illustrated example, this is not necessary and this approximate relationship may be
determined as a part of one or more other procedures, including, possibly, separate
procedures. This approximate relationship may be used as a part of the block 223 determination
of changes to the intensities of the dose-change beamlets 164. The approximate relationship
between the intensities of dose-change beamlets 164 and the magnitude of a dose change
to a dose modification voxel may have a form
D=
gni where
D is the dose (or dose change) for the dose modification voxel,
i is an approximate intensity (or intensity change) value for a dose-change beamlet
164,
n represents the number of dose-change beamlets 164 for the dose modification voxel
and g is a scaling variable. The scaling variable
g may be determined in block 47. Thus, for a particular magnitude
D of block 220 desired dose modification, the adjustments
i to the intensity values of the dose-change beamlets 164 may be determined in accordance
with this approximate relationship.
[0057] Restrictions may be applied to the block 223 intensity changes to the dose-change
beamlets 164. Such restrictions may be related to practical considerations - e.g.
to more accurately reflect dose distributions which are achievable in practice. By
way of non-limiting example, such intensity-change restrictions may include:
- the intensities of the dose-change beamlets 164 (after the block 223 changes) must
be greater than a minimum threshold (e.g. negative intensities are not permitted);
- the intensities of the dose-change beamlets 164 (after the block 223 changes) must
be less than a maximum threshold;
- the intensities of the dose-change beamlets 164 (after the block 223 changes) must
be controlled such that spatial variations are limited over the two-dimensional extent
of a corresponding dose distribution 165 or a dose distribution grid 162 (see Figure
6). For example, such a restriction may limit the maximum (magnitude or percentage)
change in intensity between immediately adjacent (or some range of adjacent) beamlets
164 in a particular dose distribution 165; and/or
- the like.
[0058] If it is determined, in block 223, that a prospective intensity change to one or
more dose-change beamlets 164 would violate an intensity-change restriction, then
a variety of strategies may be employed in block 223 to overcome such a violation.
By way of non-limiting example, such strategies may involve:
- rejecting the block 220 desired dose modification;
- adjusting the block 220 desired dose modification magnitude;
- the one or more beams 159 corresponding to dose-change beamlets 164 resulting in intensity
limit violations may be omitted from subsequent dose estimation (discussed further
below);
- the intensity changes to one or more beamlets 164 associated with secondary dose modification
voxels may be omitted or the marginal region of secondary dose modification voxels
may be reduced;
- the magnitudes of one or more intensity changes applied to one or more corresponding
dose-change beamlets 164 may be modified until the intensity-change restriction is
no longer violated;
- a combination of two or more of the above strategies; and/or
- the like.
Once it is assured that no intensity-change restrictions are violated by the prospective
block 223 intensity changes, then the intensities of the dose-change beamlets 164
are modified as discussed above.
[0059] In some embodiments, block 221 may involve identifying secondary dose-change beamlets
164 and block 223 may involve adjusting the intensity values of secondary dose-change
beamlets 164 in addition to the identification and intensity value adjustment of the
primary dose-change beamlets 164 in accordance with the techniques discussed above.
Secondary dose-change beamlets 164 may be identified in marginal regions proximate
to the primary dose-change beamlets 164. By way of non-limiting example, such secondary
dose-change beamlets 164 could be identified to be in a marginal region less than
or equal to a threshold number of beamlets away from a primary dose-change beamlet.
The threshold number of beamlets that define the marginal region may be operator-configurable,
may be a system parameter which may be determined by one or more suitable calibration
procedures, may be a system parameter which may be determined based on empirical testing
and/or data and/or the like.
[0060] The block 223 intensity adjustments to the secondary dose-change beamlets 164 may
be less than corresponding adjustments to the primary dose-change beamlets 164. For
example, the block 223 intensity adjustments to secondary dose-change beamlets 164
may be a fraction
a of the intensity adjustments to the primary dose-change beamlets (where 0<=
a<=1). This fraction may be a function of the distance between a given secondary dose-change
beamlet and its corresponding primary dose-change beamlet -e.g. within a marginal
region of 3 beamlets around a primary dose-change beamlet, the fraction
a may be relatively high for the secondary dose-change beamlets that are nearest neighbors
to the primary dose-change beamlet; lower for the secondary dose-change beamlets that
are spaced by one beamlet from the primary dose-change beamlet; and lowest for the
secondary dose-change beamlets that are spaced by two beamlets from the primary dose-change
beamlet.
[0061] In other respects (e.g. for the purposes of other procedures involved in the methods
and systems described herein), secondary dose -change beamlets 164 may be treated,
for the most part, in the same manner as primary dose-change beamlets 164. Accordingly,
both primary dose-change beamlets and secondary dose-change beamlets may be referred
to simply as dose-change beamlets.
[0062] Method 18 (Figure 4B) then proceeds to block 224 which involves determining changes
in the achievable dose distribution based on the block 223 changes to the intensities
of the dose-change beamlets 164 and updating the dose distribution accordingly. It
will be appreciated that modifying the intensities of the dose-change beamlets 164
will not only impact the estimated dose delivered to the desired dose modification
voxels identified in block 220, but will also impact the estimated dose delivered
to other voxels (e.g. voxels in and around the paths of the dose-change ray lines
163). Typically, the maximum changes to the achievable dose distribution (caused by
changes to the intensities of the dose-change beamlets 164) will occur in the desired
dose modification voxels and the changes to the achievable dose distribution will
be somewhat less in voxels surrounding the desired dose modification voxels.
[0063] The block 224 estimation of the changes in the dose distribution may involve the
use of known dose estimation techniques such as, Monte Carlo, collapsed cone convolution,
pencil beam, anisotropic analytical algorithm, Boltzman equation solvers and/or the
like. The block 224 estimation of dose distribution may involve one or more of the
rapid dose distribution estimation techniques described herein. It is generally desirable
that the block 224 dose estimation technique be computationally efficient. Block 224
may involve using dose estimation techniques that are able to update dose estimates
resulting from intensity changes to particular dose-change beamlets 164 along individual
dose-change ray lines 163 which may be more computationally efficient. The resulting
changes in the dose estimates may then be added (or subtracted in the case of a dose
reduction) to the existing dose distribution estimate. In this regard, block 224 need
only involve updating the achievable dose distribution for the dose-change beamlets
164 whose intensities were modified in block 223 - i.e. it is not necessary to recalculate
the entire dose distribution estimate in block 223. Typically, the number of dose-change
beamlets 164 having their intensities modified in block 223 will be a relatively small
subset of the beamlets 164 associated with a given beam 159. In accordance with some
dose estimation techniques, the block 224 dose estimation update may therefore consume
a relatively small amount of computational resources compared to a full dose distribution
estimate.
[0064] Figure 4D depicts a dose-estimation update method 50 for determining the changes
in achievable dose distribution which may be used in block 224 according to a particular
embodiment. Dose-estimation update method 50 of the illustrated embodiment assumes
that a plurality of dose modification voxels were specified in block 220 (e.g. the
dose distribution changes requested by an operator in block 220 resulted in modifications
to the dose at a plurality of coordinates/voxels). Dose-estimation update method 50
begins in block 51 which involves an inquiry into whether there are more dose-modification
voxels to be considered in the dose-estimation update. In the first loop through method
50, the block 51 inquiry will typically be positive (block 51 YES update path) and
method 50 will proceed to block 52. Before proceeding to block 52, block 51 may also
involve selecting one dose modification voxel to be the current dose modification
voxel for this iteration of the method 50 dose-estimation update loop.
[0065] Block 52 involves identifying the dose-change beamlets 164 which correspond to the
current dose modification voxel. Such dose-change beamlets 164 may be those identified
in block 221 (Figure 4B) as having corresponding dose-change ray lines 163 that intersect
the current dose modification voxel. For each such dose-change beamlet 164, method
50 involves obtaining the intensity change for that beamlet 164 (in block 54) and
estimating the dose contribution attributable to the intensity change for that beamlet
164 (in block 56). The block 56 estimation of the dose contribution attributable to
the intensity change for a particular dose-change beamlet 164 may involve the use
of known dose estimation techniques such as, Monte Carlo, collapsed cone convolution,
pencil beam, anisotropic analytical algorithm, Boltzman equation solvers and/or the
like. The block 56 dose estimation may additionally or alternative involve one or
more of the rapid dose distribution estimation techniques described further below.
[0066] Method 50 then proceeds to block 58 which involves summing the block 56 estimated
dose contributions for all of the dose-change beamlets 164 corresponding to the current
dose modification voxel to obtain an estimated dose distribution update for the current
dose modification voxel. In block 60, the block 58 dose distribution update for the
current dose modification voxel may optionally be scaled to provide a scaled dose
update grid corresponding to the current dose modification voxel. The term dose update
grid may be used interchangeably with dose update distribution or dose modification
distribution. As discussed above, an approximate relationship may be established (e.g.
in block 47 of initialization method 20) between the dose change to a dose modification
voxel and the corresponding intensity changes to the dose-change beamlets 164 and,
in accordance with this approximate relationship, changes to the intensities of the
dose-change beamlets 164 may be established in block 223. However, this relationship
is only approximate. Consequently, the block 58 sum of the dose change contributions
for each of the dose-change beamlets 164 may not yield the desired magnitude (e.g.
the block 220 desired magnitude) of dose change to the dose modification voxel. In
such cases, the block 58 sum of the dose change contributions for each of the dose-change
beamlets 164 may be scaled in block 60 to obtain a scaled dose update grid which achieves
the desired magnitude of dose change (e.g. the block 220 desired dose change magnitude).
The block 60 scaling of the dose update grid may also be accompanied by corresponding
scaling to the intensities of the dose-change beamlets. The dose-change beamlets may
be scaled by a similar factor. For example, if the block 60 scaling of the dose change
grid involves a scaling factor s, then the intensities of the dose-change beamlets
may be scaled by the same scaling factor s.
[0067] Since there may be an approximate relationship established between the dose change
to the dose modification voxel and the corresponding intensity changes to the dose-change
beamlets 164, the block 60 scaling may be minimal. In some examples, block 60 scaling
is not used. In some instances, even where block 60 scaling is used, it may be desirable
to limit the amount of block 60 scaling. For example, it may be undesirable to scale
in a manner which may result in one or more beamlet intensities that violate beam
restrictions, such as any of the beam restrictions discussed above. In some cases
(e.g. because scaling is not used or is limited in amount), it may not be possible
to achieve the desired dose change magnitude (e.g. the block 220 dose modification
magnitude) in the dose modification voxel. This circumstance is permissible.
[0068] Figure 12 schematically depicts an example of a scaled dose update grid 234 corresponding
to a particular dose modification voxel 231. In the Figure 12 depiction, the current
dose modification coordinate is voxel 231 in target tissue structure 153. Referring
to Figures 4B, 4D and 12, block 221 identifies the illustrated ray lines 163 (and
corresponding beamlets 164) for a series of beams 159 to be the dose-change ray lines
163 and dose-change beamlets 164. The changes to the intensity values of the corresponding
dose-change beamlets 163 (determined in block 223 of Figure 4B) are obtained in each
iteration of block 54 (Figure 4D) and their respective dose distribution contributions
are estimated in each iteration of block 56 (Figure 4D). These block 56 dose distribution
contributions are summed in block 58 and optionally scaled in block 60 to obtain scaled
dose update grid 234 illustrated in Figure 12.
[0069] Figure 12 shows a dose-update grid 234 comprising dose update values. The dose update
values determined for particular voxels in grid 234 are represented by the numerical
values shown in the boxes of dose-update grid 234. The dose update values for grid
234 are expressed in terms of scaled dose update values - i.e. the dose update value
of 100% at the current dose modification voxel 231 represents a dose update of 100%
of the desired dose change magnitude (e.g. the block 220 desired dose change magnitude)
and the dose update values of the other voxels around current dose modification voxel
231 represent lower percentages of the desired dose change magnitude. Figure 12 shows
that the maximum dose distribution change occurs at the current dose modification
voxel 231 but that dose distribution changes also occur to a lesser extent in surrounding
voxels. It will be appreciated that dose-update grid 234 shown in Figure 12 is a two-dimensional
representation, but that in practice the dose-update grid determined in accordance
with method 50 will be three-dimensional.
[0070] The block 60 dose-update grid 234 may represent an amount of dose to add to (or subtract
from) an overall achievable dose distribution in block 64. As discussed above, an
initial overall achievable dose distribution may be determined in block 46 (Figure
4C). The block 46 initial estimate of the overall dose distribution may be updated
in one or more previous iterations of method 18 (Figure 4B). Block 64 may involve
adding the dose update values of dose-update grid 234 to the overall dose distribution
estimate. The overall dose distribution estimate will then be updated with the dose
distribution contribution for the current dose modification voxel 231. Method 50 then
loops back to block 51 to determine whether there are other dose modification voxels
to be considered in the method 50 dose-estimation update. After one or more loops
through blocks 52-64, the block 51 inquiry will be negative, terminating method 50.
[0071] It will be appreciated by those skilled in the art that dose-estimation update method
50 of the Figure 4D represents one particular method of updating dose estimates as
a part of block 223 (Figure 4B). In other examples, equivalent dose-estimation updates
could be determined using different orders of operations. By way of non-limiting example,
rather than looping through dose modification voxels and dose-change beamlets, the
same or similar dose-estimation update results could be obtained by identifying all
dose modification voxels, summing the intensity changes to the dose-change beamlets
resulting from all of the dose modification voxels and then estimating the dose changes
corresponding to all of the dose-change beamlets in a single dose estimation process.
Other orders of operations are conceivable for estimating dose updates as a part of
block 224 (Figure 4B). In some examples, the scaling and multiplication operations
of blocks 60 and 62 are not required as the magnitude of the desired dose change for
a particular dose modification voxel (see block 220 of Figure 4B) may already be taken
into account in the amount of the intensity changes to the dose-change beamlets 164.
In some examples, where all dose modification voxels are simultaneously updated, scaling
may occur at some level that is a combination (e.g. some type of average) of the desired
scaling for all of the dose modification voxels.
[0072] Method 50 involves obtaining individual intensity changes to dose-change beamlets
(in block 54), estimating dose changes (in blocks 56, 58, 60, 62) and then adding
(or subtracting) dose changes to the existing dose distribution (in block 62) to obtain
the block 224 (Figure 4B) updated dose distribution. This works with changes to intensities
and doses and is possible because many dose estimation techniques obey the principal
of superposition. In other examples, dose-estimation update methods which may be used
in block 224 of method 18 (Figure 4B) may involve discarding an existing dose distribution
(or part(s) thereof) and then estimating a replacement dose distribution (or replacement
part(s) thereof). Such replacement dose distributions (or replacement part(s) thereof)
may be based on the absolute values of the adjusted intensities of the dose-change
beamlets 164 (i.e. rather than the changes to the intensities of the dose change beamlets
164).
[0073] In one example the parts of an existing dose distribution which could be discarded
comprise the parts of the existing dose distribution contributed by the previous values
of dose-change beamlets 164, in which case the estimated replacement parts of the
updated dose distribution estimate would be the dose contributions from the new intensity
values of the dose-change beamlets 164. In another example the parts of an existing
dose distribution which could be discarded comprise the parts of the existing dose
distribution contributed by the intensity distributions 165 of beams 159 (e.g. intensity
distributions 165 having one or more dose-change beamlets 164), in which case the
replacement parts of the updated dose distribution estimate would be the dose contributions
from the modified intensity distributions 165 of beams 159 (e.g. intensity distributions
165 modified by updated intensity values for one or more dose-change beamlets 164).
In another example, the entire existing dose distribution could be discarded and a
replacement dose distribution could be estimated based on dose contributions from
all of the updated intensity distributions 165 of all of the beams 159. These techniques
of discarding parts of existing dose distributions and estimating replacement parts
for the updated dose distribution estimate may be particularly useful where they are
used in conjunction with the convolution technique for rapid estimation of achievable
dose distribution which is discussed further below.
[0074] At the conclusion of method 50 and/or block 224 (Figure 4B), the achievable dose
distribution has been updated to accommodate the desired block 220 dose distribution
changes (e.g. operator-requested dose distribution changes). Those skilled in the
art will appreciate from the discussion above, that there are a variety of techniques
in which to update the achievable dose distribution in block 224. While some such
techniques have been discussed above, there may still be other suitable techniques
which may be modifications of those discussed above or which may be different from
those discussed above. Nevertheless, at the conclusion of block 224, the achievable
dose distribution has been updated to accommodate the desired block 220 dose distribution
changes. The updated estimate of the dose distribution at the conclusion of block
224 may be displayed for the operator. Returning to Figure 4B, method 18 may optionally
proceed to block 225 where one or more dose quality metrics may be updated on the
basis of the block 224 updated dose distribution. Such dose quality metrics may also
be displayed for the operator.
[0075] Figure 7B shows a cross-section of the updated dose distribution estimate 173 (determined
in block 224) as a result of the Figure 7A desired dose change obtained or otherwise
determined in block 220. Recalling that the Figure 7A dose change corresponded to
a desired reduction in the dose delivered to voxel 171 of healthy tissue structure
152, Figure 7B shows that the maximum dose reduction occurs at voxel 171. However,
updated dose distribution estimate 173 of Figure 7B also shows that the dose estimate
was reduced for voxels surrounding voxel 171. The updated dose distribution shown
in Figure 7B may be displayed to an operator in block 224.
[0076] Figure 9B shows a cross-section of the updated dose distribution estimate 193 (determined
in block 224) as a result of the Figure 9A desired dose change determined in block
220. As discussed above, the Figure 9A desired dose change corresponds to a desired
change in the Figure 8A DVH which involves an increase in the dose delivered to target
tissue structure 153 and corresponding increases to the dose delivered to dose modification
voxels 191. Figure 9B shows that the maximum dose increases occur at dose modification
voxels 191, but that dose increases also occur for voxels surrounding voxels 191.
The updated dose distribution shown in Figure 9B may be displayed to an operator in
block 224. Figure 8B shows the updated DVHs corresponding to the updated dose distribution
estimate of Figure 9B. The Figure 8B DVHs may be determined and displayed to an operator
in block 225. Figure 8B shows that the target tissue DVH 182 changes most near the
desired change point 181, but that smaller changes also occur to healthy tissue DVH
186.
[0077] Figure 11B shows a cross-section of the updated dose distribution estimate 213 (determined
in block 224) as a result of the Figure 11A desired dose change determined in block
220. As discussed above, the Figure 11A desired dose change corresponds to a desired
change in the Figure 10 NTCP biological index which involves a decrease in the dose
delivered to healthy tissue structure 152 and corresponding decreases to the dose
delivered to dose modification voxels 211. Figure 11B shows that the maximum dose
decreases occur at dose modification voxels 211, but that dose decreases also occur
for voxels surrounding voxels 211. The updated dose distribution shown in Figure 11B
may be displayed to an operator in block 224. Although not expressly shown in the
drawings, updated biological indices may be determined and displayed to an operator
in block 225.
[0078] In some embodiments, method 18 and/or portions thereof may be performed in loops.
For example, at the conclusion of blocks 224 and/or 225, method 18 may loop back to
block 220 to permit additional dose changes (e.g. operator input of additional desired
dose changes). In some embodiments, dose estimation updates (block 224) and/or updates
to dose quality metrics (block 225) may be performed and/or displayed periodically.
The periods between such computation and/or display updates (which need not be temporal
periods) may be defined using a variety of techniques. By way of non-limiting example,
updates may be performed:
- each time that a block 220 desired dose modification is requested or otherwise determined;
- after multiple block 220 desired dose modifications are requested or otherwise determined;
- after a time interval;
- if several block 220 desired dose modifications are requested, it may be desirable
to display a result which includes some subset of the requested dose modifications;
- after an achievable dose distribution and/or dose quality metric has changed by a
threshold amount;
- a combination of the above (e.g. after a time interval or a threshold number of desired
dose modifications are requested); and/or
- the like.
[0079] Some of the procedures of method 18 may overlap with one another. For example, an
operator may request multiple desired dose modifications (block 220) prior to the
completion of the remainder of method 18. As desired dose modifications are communicated
by the operator (or otherwise obtained in block 220), the rest of method 18 may be
carried out, so that modifications may be continuously applied and the achievable
dose distribution may be continuously updated. It may occur that one or more further
desired dose modifications are requested prior to completion of the rest of method
18 for a previous desired dose modification update. Further desired dose modifications
may be places in a queue so that, once method 18 is completed for a particular desired
dose modification, method 18 may be completed for the next desired dose modification
in the queue. In this way all desired dose modification request changes will eventually
be processed. In other examples, further desired dose modification requests may only
be permitted after method 18 has completed for a previous desired dose modification
request. Additional example schemes for addressing further desired dose modification
requests while method 18 is being carried out for previous desired dose modification
requests include, but are not limited to, rejection of every 2nd, 3rd or
Nth request (where N is an integer) while the remaining requests are placed in a queue.
[0080] During method 18 (e.g. as a part of block 224 and/or method 50 (Figure 4D)), it may
be desirable to prohibit or restrict certain desired dose modification requests (e.g.
block 220 desired dose modification requests). For example, reducing dose to target
tissue structures 153 below a threshold or increasing dose to healthy tissue structures
152 above a threshold may be considered undesirable. Such thresholds may comprise
operator-specified thresholds, system threshold parameters and/or the like, for example.
Restrictions (e.g. thresholds) may be specified for the dose distribution estimate
itself and/or for one or more dose quality metrics. As part of method 18, proposed
updates to the achievable dose distribution and/or proposed updates to one or more
dose quality metrics may be evaluated with respect to one or more such restrictions.
In the event that such a restriction is violated, a variety of actions may be taken.
By way of non-limiting example:
- the block 220 desired dose modification request corresponding to the restriction violation
may be rejected;
- the magnitude of the block 220 desired dose modification request may be adjusted so
that the restriction is no longer violated;
- one or more additional changes to the block 220 desired dose modification request
may be made in attempt to overcome the violation of the restriction; and/or
- the like.
[0081] In the event that one or more additional changes to the block 220 desired dose modification
request are changes in attempt to overcome violation of the restriction, the following
exemplary procedure may be used:
- restriction violating coordinates (e.g. voxels) corresponding to the voxels in the
dose distribution that violate the restriction may be determined. Such restriction
violating coordinates may be determined in a manner similar to the determination of
the block 220 dose modification coordinates (e.g. for dose quality metrics).
- dose modifications may be applied to the restriction violating coordinates and a resulting
dose distribution estimates may be obtained in accordance with method 18 of Figure
4B.
- after the new dose distribution estimates are calculated for the changes to the restriction
violating coordinates, the new dose distribution estimates may be evaluated for restriction
violations.
- If one or more restrictions remains violated, then the procedure can be repeated until
there are no longer restriction violations.
[0082] At the conclusion of method 18 (any loops or any portions thereof), method 18 yields
an achievable dose distribution and/or an estimated dose quality metric. Preferably,
such achievable dose distribution and/or estimated dose quality metric will meet the
operator's treatment objectives. The achievable dose distribution and/or dose quality
metrics may be output for use by another method or system. Such methods or systems
may comprise, for example, a computerized database, a treatment plan optimization
system, a radiation delivery apparatus and/or as an input to any other system or device
used in radiation treatment. The corresponding beamlet intensities and/or dose restrictions
associated with the Figure 18 dose distribution estimation may also be output for
use by any such method or system.
[0083] By way of non-limiting example, as discussed above, method 18 of Figure 4B may comprise
a method for implementing the block 142 generation and manipulation of achievable
dose distribution as a part of radiation delivery method 16 (Figure 4A). In such an
example, the block 224 dose distribution estimates and the corresponding beamlet intensities
may be used in block 143 of radiation delivery method 18 to perform an iterative optimization
process or the like to derive radiation delivery parameters (e.g. beam configuration
parameters and/or beam delivery parameters). If iterative optimization is used in
block 143, the dose estimation process involved in such optimization may (but not
necessarily) be the same as the dose estimation process involved in the dose generation
and manipulation procedure of block 142 (e.g. in block 46 of method 20 and block 56
of method 50). For example, in some examples, the dose estimation procedures used
in dose generation and manipulation block 142 may comprise one or more of the rapid
dose estimation techniques described in more detail below; however, the block 143
iterative optimization may involve one or more traditional methods of dose estimation
(e.g. Monte Carlo, collapsed cone convolution, pencil beam, anisotropic analytical
algorithm, Boltzman equation solvers and/or the like) which may be more accurate for
determining radiation delivery parameters.
[0084] If iterative optimization is used in block 143, one or more outputs from the block
142 (method 18) generation and manipulation of estimated dose may be used to aid in
the derivation of the block 143 radiation delivery parameters. By way of non-limiting
example:
- the block 224 achievable dose distribution and/or block 225 estimated dose quality
metric(s) may be used in block 143 to define the optimization objectives (e.g. cost
function or the like) of the optimization process;
- the beamlet intensities determined in block 223 and corresponding to the final block
224 achievable dose distribution may be used to determine the beam intensity required
of the radiation delivery apparatus;
- other metrics derived from the block 224 achievable dose distribution may be used
to define optimization objectives (e.g. a cost function or the like) of the block
143 optimization process; and/or
- the like.
[0085] Block 143 does not necessarily require the performance of an optimization process.
In some examples, the output(s) of the block 142 generation and manipulation of achievable
dose distributions (e.g. block 224 achievable dose distributions, block 225 dose quality
metrics and/or block 223 beamlet intensities) may lead be convertible directly to
radiation delivery parameters of sufficient accuracy. Such direct derivation of radiation
delivery parameters (i.e. without iterative optimization) in block 143 may occur,
for example, where the block 224 achievable dose distributions are calculated according
to a sufficiently accurate estimation technique and various restrictions (e.g. on
beamlet intensities and/or dose estimates) are sufficiently robust.
[0086] Radiation treatment method 16 may then proceed to block 144, where the block 143
radiation delivery parameters may be transferred to a radiation delivery apparatus.
In block 145, a controller associated with the radiation delivery apparatus (equipped
with the radiation delivery parameters) may then cause the radiation delivery apparatus
to deliver radiation to the subject. The radiation received by the subject is preferably
similar to the achievable dose distribution output predicted in blocks 142 (Figure
4A) and/or 224 (Figure 4B).
[0087] Systems and methods according to various examples described herein involve estimating
dose distributions based on one or more beamlet intensities. Non-limiting examples
of estimating dose distributions include: the block 218 initialization procedure of
method 18 (Figure 4B) and the corresponding estimate of the initial dose distribution
in block 46 of initialization method 20 (Figure 4C); the block 224 dose-estimation
update procedure of method 18 (Figure 4B) and the corresponding estimation at block
56 (Figure 4D); and any dose distribution estimation that may take place in the block
143 procedure for deriving radiation delivery parameters (e.g. during iterative optimization).
As discussed above, there are a variety of known techniques (e.g. Monte Carlo, collapsed
cone convolution, pencil beam, anisotropic analytical algorithm, Boltzman equation
solvers and/or the like) for estimating dose distribution based on beamlet intensities.
Any such techniques could be used in any of the intensity distribution estimation
procedures described herein, although it may be preferable that one technique is used
consistently throughout block 142 and one technique is used consistently throughout
block 143 of radiation treatment method 16.
[0088] One aspect of the disclosure provides different methods for estimation of achievable
dose distributions. Such methods may be used to perform the dose distribution estimation
procedures in any of the other methods and systems described herein. Methods for estimating
achievable dose distributions are provided which are relatively rapid in comparison
to currently available dose estimation techniques, such as Monte Carlo, collapsed
cone convolution, pencil beam, anisotropic analytical algorithm, Boltzman equation
solvers and/or the like. Dose distributions estimated in accordance with the inventive
methods described herein may be referred to as rapid dose distribution estimates to
contrast them with traditional dose distribution estimates obtained using known techniques.
Methods of estimating achievable dose distributions according to various embodiments
of the disclosure may involve simplifications based on ray lines 163 (and corresponding
beamlets 164 of intensity distributions 165) emanating from radiation source 161 and
knowledge of how such ray lines 163 interact with calculation grids which are used
to map three-dimensional space in subject 160 (see Figure 6). By way of non-limiting
example, such simplifications may involve:
- using simplified models of radiation scatter and/or radiation transport as compared
to traditional dose distribution estimation techniques;
- omitting the effects of inhomogeneous subject density;
- omitting attenuation of beams 159 as a function of depth in subject 160;
- ignoring the distance from the radiation source to the subject 160;
- a combination of the above; and/or
- the like.
[0089] Given a set of beams 159 with known intensity distributions 165, rapid estimates
of achievable dose distributions (and corresponding dose quality metrics) determined
in accordance with some of the inventive methods described herein may not yield the
same achievable dose distributions (and corresponding dose quality metrics) as traditional
dose estimation methods. However, where suitable limits are imposed on beams 159,
corresponding intensity distributions 165 and the intensities of individual beamlets
164, rapid dose distribution estimation techniques described herein may yield dose
distribution estimates that are reasonably close to those that are physically deliverable.
Suitable examples of limitations on beams 159, corresponding intensity distributions
165 and the intensities of individual beamlets 164 are described above. In some examples,
where rapid estimation of achievable dose distributions are used during the process
of treatment planning and/or delivery (e.g. method 16 of Figure 4A), iterative optimization
procedures may be used (e.g. in block 143 of method 16) to determine radiation delivery
parameters (e.g. physical system parameters) capable of delivering the desired fast
dose distribution estimates.
[0090] Figure 13 is a schematic illustration of a method 70 for rapid estimation of achievable
dose distribution according to a particular embodiment. As discussed in more detail
below, method 70 for rapid estimation of achievable dose distributions involves, for
each beam 159 in a beam configuration, convolving the two-dimensional intensity distribution
165
i(x,y) associated with the beam 159 with a two-dimensional dose estimate kernel
k(x,y) to obtain a convolved intensity distribution and, for each convolved beamlet in the
convolved intensity distribution, projecting the convolved intensity value along the
ray line 163 associated with the convolved beamlet. Estimating achievable dose distributions
using a dose estimate kernel
k(x,y) in accordance with method 70 may be an efficient way of estimating an achievable
dose distribution based on an intensity distribution.
[0091] Rapid dose distribution estimation method 70 commences in block 71 which involves
an inquiry into whether there are more beams to be considered in the rapid dose distribution
estimation. In the first loop through method 70, the block 71 inquiry will typically
be positive (block 71 YES update path) and method 70 will proceed to block 72. Before
proceeding to block 72, block 71 may also involve selecting one beam to be the current
beam for this iteration of the method 70 rapid dose distribution estimation loop.
In block 72, method 70 involves performing a convolution operation which comprises
convolving the two-dimensional intensity distribution
i(x,y) associated with the current beam with a two-dimensional dose estimate kernel
k(x,y) to obtain a two-dimensional convolved intensity distribution f
(x,y). The coordinates x,y may be defined in the plane of grid 162 of the intensity distribution
165 (see Figure 6). The resultant convolved intensity distribution
f(x,y) may be defined for, or otherwise mapped to, the coordinates associated with the beamlets
164 of the current beam which may, in the context of the convolved intensity distribution
f(x,
y), be referred to as convolved beamlets 164. As part of block 72, each convolved beamlet
164 is associated with a corresponding ray line 163 and a corresponding convolved
intensity value.
[0092] The dose estimate kernel
k(x,y) may be intended to approximate the amount of radiation scatter and energy transport
resulting from radiation interacting with tissue. In some examples, the dose estimation
kernel
k(x,
y) comprises a point spread function. In some examples, the dose estimation kernel
k(x,y) comprises a linear combination of a plurality of point spread functions. In one exemplary
example, the dose estimate kernel
k(x,y) comprises a linear combination of one or more 2-dimensional Gaussian functions:
where
Ai are magnitude variables of the various Gaussian functions and
σi are variables representative of the radial spread of the various Gaussian functions.
The variables
Ai and
σi may be operator-configurable, may be system parameters which may be determined by
one or more suitable calibration procedures, may be system parameters which may be
determined based on empirical testing and/or data and/or the like. The variables
Ai and
σi may depend on the type of radiation. It will be appreciated by those skilled in the
art that the equation (1) point spread function is merely an example point spread
function and that dose estimation kernel
k(x,y) may comprise a variety of other point spread functions and/or linear combinations
of point spread functions. In some examples, the dose estimation kernel
k(x,y) (or one or more parameters thereof) may be experimentally determined (e.g. from calibration
type measurements and/or the like). In some examples, the dose estimation kernel
k(x,y) (and/or its parameters) may be stored in accessible memory (e.g. in a look up table
or the like).
[0093] Convolution operations can be computationally intensive and can consume relatively
large amounts of processing resources. To reduce this burden on computational resources,
the block 72 convolution may involve converting the two-dimensional intensity distribution
i(x,y) and the two-dimensional dose estimate kernel
k(x,y) to the Fourier domain. Advantageously, a convolution operation in the spatial domain
may be implemented as a multiplication operation in the Fourier domain. The two-dimensional
intensity distribution
i(x,y) and the two-dimensional dose estimate kernel
k(x,y) may be converted to the Fourier domain using any of a wide variety of known computational
techniques for performing Fourier transforms (e.g. fast Fourier transforms (FFT) and/or
the like). The convolved intensity distribution
f(x,y) may therefore be calculated according to:
where
FT[●] and
IFT[●] are respectively Fourier transform and inverse Fourier transform operators. It will
be appreciated that, in absence of a change to the kernel function
k(x,y), the Fourier transform of the kernel function
FT[k(x,y)] need only be calculated once and the result may be stored (e.g. in a look up table
in accessible memory and/or the like).
[0094] Method 70 then proceeds to block 74 which involves a loop for each convolved beamlet
164 in the current beam 159. For each convolved beamlet 164 in the current beam159:
block 76 involves projecting the ray line 163 corresponding to the convolved beamlet
164 into the calculation grid and identifying voxels in the calculation grid which
are intersected by the ray line 163 and block 78 involves adding the convolved intensity
value for the convolved beamlet 164 to each voxel identified in block 76. Method 70
then loops back to block 71 to determine whether there are more beams to be considered
in the method 70 rapid dose estimation. After one or more loops through blocks 72-78,
the block 71 inquiry will be negative, terminating method 70.
[0095] Rapid dose distribution estimation method 70 involves the principal of superposition.
As discussed above in connection with dose estimation update method 50, dose estimation
techniques which involve the principal of superposition may operate on changes (e.g.
determining changes to dose distributions that result from changes to beam and/or
beamlet intensities) or on absolute values (e.g. discarding dose contributions from
previous values of beam and/or beamlet intensities and estimating new dose contributions
based on the new absolute values of the beam and/or beamlet intensities).
[0096] The above-described method 70 for rapid dose distribution estimation may be augmented
by incorporating an estimate for attenuation of radiation as it passes through the
subject. For this purpose, in some examples, an additional attenuation function
a(d) may be applied to account for such attenuation. Such an attenuation function
a(d) may cause the block 72 convolved intensity values to decrease with distance
d along the ray lines 163 which they are projected in block 76, so that their respective
block 78 dose contributions decrease with distance
d along their respective ray lines 163. By way of non-limiting example, an attenuation
function
a(d) may be multiplied to the dose contribution values in block 78 before such dose contribution
values are added to each voxel intersected by ray lines 163. A variety of different
decreasing attenuation functions
a(d) are suitable to model this attenuation and the choice of particular attenuation function
a(d) may depend on the characteristics of the radiation. In one particular example, an
attenuation function
a(d) may be provided by an exponential function of the form:
where
B is a magnitude variable,
k is a variable characterizes the rate of attenuation with depth and
d represents the distance along a particular ray line 163. In some examples, the variable
d may represent the depth of penetration into the body of the subject - i.e. there
is negligible attenuation prior to the radiation impinging on the body of the subject.
The variables
B and
k may be operator-configurable, may be system parameters which may be determined by
one or more suitable calibration procedures, may be system parameters which may be
determined based on empirical testing and/or data and/or the like. Like the Fourier
transform of the kernel function, the values of the attenuation factor (e.g. equation
3) may be pre-calculated and stored in a look up table or the like.
[0097] In some examples, beam configurations are contemplated which involve a plurality
of beams 159 at a corresponding variety of locations along a trajectory which involves
a 360° rotation of the radiation source with respect to the subject. In some examples,
it is contemplated that the radiation source will move continuously with respect to
the subject about the 360° trajectory. In such examples, the 360° trajectory can be
approximated by a plurality of sample beams. For the purpose of this description,
sample beams on a trajectory where it is contemplated that the radiation source will
move continuously relative to the subject can be treated in the same manner as discrete
beams 159.
[0098] Figure 14 schematically depicts a 360° trajectory 240, a plurality of beams 159 at
angularly evenly spaced locations around trajectory 240 and a two-dimensional cross-section
of a dose update grid 242 estimated to be achievable by the plurality of beams 159.
Beams 159 may be discrete beams or sample beams associated with a continuously moving
radiation source. The achievable dose update grid 242 (having voxel dose contributions
represented by the numbers in the boxes of the illustrated grid) has been scaled to
a value of 100 at the dose modification voxel 241. The Figure 14 dose update grid
242 may be obtained, for example, in accordance with block 60 of method 50 (Figure
4D). The Figure 14 dose update grid results from an equal intensity change to each
of the beamlets associated with the ray lines 163 that intersect dose modification
voxel 241.
[0099] Figure 15 depicts an example profile 243 of the Figure 14 achievable dose update
grid 242 that intersects dose modification voxel 241. The peak of the dose change
depicted in example profile 243 occurs at the location of dose modification voxel
241 (x=0 in Figure 15). Due to the 360° trajectory 240 and plurality of beams 159,
the magnitude of the dose changes in areas surrounding dose modification voxel 241
decreases rapidly with distance from dose modification voxel 241. To achieve this
property (i.e. where relatively small changes occur to the dose at locations away
from particular dose modification voxels 241), it may be desirable that there be a
relatively large number of beams 159 in 360° trajectory 240. As discussed above, in
some examples, the radiation source moves continuously with respect to the subject
and beams 159 are actually sample beams. In other examples, the angular separation
between discrete beams 159 along the 360° trajectory 240 is less than 20°. In some
examples, this angular separation is less than 15°.
[0100] An additional property of the 360° trajectory 240 and the corresponding dose update
grid 242 is that the same dose update grid may be translated to any desired dose modification
voxel. An example of this translatability is shown in Figure 16 which depicts a dose
update grid 244 which is the same as dose update grid 242 (Figure 14) except that
dose update grid 244 of Figure 16 has been translated to a different dose modification
voxel 245. The translatability of the dose update grid 242 for the 360° trajectory
240 may be permitted because every voxel in the plane of motion of the radiation source
along trajectory 240 has a full 360° of ray lines which intersect the voxel and corresponding
beamlets within beams 159. Some examples may achieve computational efficiency by determining
the dose update grid 242 for a 360° trajectory 240 and storing the dose update grid
242 in accessible memory or the like (e.g. in a look up table). In this manner, the
dose update grid 242 for a 360° trajectory 240 may be recalled from memory each time
that it is used.
[0101] A 360° trajectory 240 is used as an example for illustration purposes. Beam configurations
with other trajectories (e.g. other angular rotation ranges and/or other) motions
may also be used. Furthermore, it may be desirable to use different angular ranges
of beams when permitting manipulation of achievable dose (e.g. in block 142). For
angular ranges of beams smaller than 360°, the corresponding angular range of ray
lines 163 intersecting a dose distribution coordinate (e.g. a dose modification voxel)
will be reduced. In some examples, dose update grids for alternative motion ranges
may also be pre-calculated and stored (e.g. in a look up table in an accessible memory)
for subsequent retrieval. A library of dose update grids referred to as radial modification
dose (RMD) distributions may be stored in accessible memory. Such a library of RMDs
may include RMDs for N incremental motion ranges from 0°-
N°, for example, where
N has incremental values going from 0°-360°. Specific angular ranges of beams may be
determined by subtracting two dose modification distributions stored in the library.
For example:
Where
θ ⇒
φ indicates an angular range of beams going from angle
θ to angle
φ. Using a subtraction operation like that of equation (4) reduces the number of actual
RMDs that must be stored in memory. A graphical example is shown in Figures 18A, 18B,
18C, where RMD dose distributions are displayed as pixel map images, in which brighter
pixels indicate higher doses. Figure 18A represents a RMD for the angular range 0°-90°
and Figure 18B represents a RMD for 0°-45°. Both the Figure 18A and 18B RMDs may be
stored in accessible memory. Assuming that it is desired to determine a RMD for the
angular range 45°-90°, then the 0°-45° RMD (Figure 18B) may be subtracted from the
0°-90° RMD (Figure 18A) to yield the 45°-90° RMD shown in the representation of Figure
18C.
[0102] Figure 17 schematically illustrates a system 300 according to a particular example.
System 300 (or portions thereof) may be used to perform some or any of the methods
described above. System 300 comprises a radiation delivery apparatus 302 which may
be similar to any of the radiation delivery apparatus described herein. Operator 312
may interact with radiation delivery apparatus 302 via user interface 302B which may
include both hardware components and software components. In the illustrated example,
the operation of radiation delivery apparatus 302 is controlled by a controller 302A.
In the illustrated example, controller 302A receives radiation delivery parameters
from an external processor 304. Processor 304 may comprise any suitable computer or
digital processing unit(s). Processor 304 may carry out computer readable software
instructions. Processor 304 may be in communication with radiation delivery apparatus
302 using any suitable communication hardware and/or software including network communication
hardware and/or software. An operator 312 may interact with processor 304 using a
suitable user interface 306 which may include both hardware components and software
components.
[0103] Processor 304 of the illustrated example has access to a computer-readable memory
308 which may house software instructions 310. In other examples, processor 304 may
obtain instructions 310 from one or more other sources. When executed by processor
304, software instructions 310 may cause processor 304 to perform one of more of the
methods described herein (e.g. radiation delivery method 16 (Figure 4A), method 18
for generation and manipulation of achievable dose (Figure 4B), dose estimation update
method 50 (Figure 4D), rapid dose estimation 70 (Figure 13) and/or the like). In the
illustrated example, processor 304, user interface 306 and memory 308 are part of
a computer system 314, although this is not necessary. In other examples, these components
(or parts thereof) may be separately implemented.
[0104] In still other examples, one or more of the methods described herein may be performed
by controller 302A (or some other suitable processor) that is part of radiation delivery
apparatus 302. In such examples, radiation delivery apparatus 302 may comprise (or
otherwise have access to) suitable memory which may comprise suitable computer software
instructions.
Examples
[0105] The examples set out below represent non-limiting examples of methods, systems and
various features of methods and systems. These non-limiting examples are for illustrative
purposes only and are not intended to represent limiting features unless otherwise
claimed below.
Pseudocode Example
[0106] Setup Initialization
- Import subject image and segmented tissue structure information (e.g. tumor target,
healthy tissue structures (e.g. organs), other tissue of interest, target prescription)
- Select dose estimate kernel (e.g. point spread function (PSF)) based on beam energy
and associated scattering characteristic(s)
- Calculate Fast Fourier Transform (FFT) of PSF and store in accessible memory for later
use
- Create voxel matrix (calculation grid) overlaid with image data from subject including
tissue structure information
- Identify voxels located within any structure(s) of interest
∘ identify voxels partially within a structure (e.g. at the edge of the structure)
and assign partial volume values (e.g. in a range of 0-1) to those voxels for the
purpose of calculating dose quality metrics based on such voxels
- Select beam configuration (e.g. number, location and direction of beams)
∘ Desired beam configuration may include a number of fixed beams (e.g. 5 to 9 beams
or other numbers of fixed beams)
∘ If desired beam configuration involves continuously moving beams, then a number
of beam samples (e.g. 10 or more beam samples) along the trajectory may be used to
represent dynamic source motion
- Initialize beamlet intensities
- Calculate estimated dose distribution resulting from initial beamlet intensities
- Calculate initial dose volume histogram (DVH) (and/or other initial dose quality metric(s))
- Display initial DVH(s) and/or initial dose distribution overlaid with image data from
subject including tissue structure information
∘ dose may be represented via cross sectional view(s) of the image data
▪ dose can be represented as a color wash
▪ dose can be represented as an iso-line plot (line contours of constant dose)
▪ dose can be represented using any other suitable display technique
∘ three-dimensional dose display (so-called "dose cloud" displays) could additionally
or alternatively be used
Beamlet Initialization
[0107]
- Create two-dimensional beamlet matrix for each beam
- For each target structure:
∘ Target structure is projected onto the beamlet matrix
∘ A margin may be added to the projection to ensure proper coverage of the target
structure by the resulting dose distribution
∘ Assign an intensity value to beamlets corresponding to target structure. Such intensity
value may be correlated with (e.g. proportional to) the prescription dose for the
target structure
- Optional:
∘ For target structures that overlap in a given beam projection, weight the beamlet
intensities in the overlapping area with the overlapping target structure that has
the highest prescription dose; or
∘ For target structures that overlap in a given beam projection, weight the beamlet
intensities in the overlapping area with the overlapping target that has the lowest
prescription dose
Rapid Estimation of Achievable Dose Distribution
[0108]
- Estimate dose contribution - For each beam:
∘ Convolve beamlet intensity matrix by the dose estimate kernel (e.g. PSF)
▪ Perform convolution by: determining the FFT of the beamlet matrix; multiplying by
the pre-calculated FFT of the dose estimate kernel; and determining the Inverse FFT
of the result
- Determine intersection of ray lines from beamlet matrix with voxels - For each voxel
where the dose has not been previously determined:
∘ For each beam:
▪ Determine the location where any ray line passing through the voxel intersects with
the beamlet matrix
▪ Store the beamlet intersection location in accessible memory for use in subsequent
dose calculations (so that the beamlet intersection location only needs to be determined
once)
▪ Achievable dose estimation may be limited to voxels of interest. For example:
• For DVH calculation, achievable dose estimation may be limited to voxels within
target and healthy tissue structure(s)
• In the case where a dose plane cross section is being displayed, achievable dose
estimation may be limited to voxels in that plane
- Calculate total dose - For each voxel:
∘ Extract the dose contribution - For each beam:
▪ Extract the convolved intensity value from the convolved intensity distribution
at the location of the previously determined beamlet matrix intersection
∘ Sum the contribution from each beam
Permit Dose Volume Histogram Manipulation
[0109]
- Operator selects a point on a DVH curve of a particular structure and requests a change
of the curve shape at that point
∘ Identify the dose and volume coordinate values of the selected point (D_selected, V_selected)
∘ The request may be communicated by selecting the DVH with a mouse or similar computer
pointing device click and then moving the mouse or similar computer pointing device
▪ movement left and/or down indicates may indicate a dose reduction (desired dose
change is negative)
▪ movement right and/or up may indicate a dose increase (desired dose change is positive)
∘ Determining a magnitude of the desired dose change may involve, for example:
▪ A fraction of the D_selected value
▪ A fraction of the maximum dose for that structure
▪ A fraction of a prescription dose assigned to that structure
▪ A fraction of the maximum prescription dose assigned to all structures
▪ A fraction correlated with (e.g. proportional to) the amount of mouse or similar
computer pointing device movement
▪ A quantity set by the operator
▪ A fixed quantity which may be an operator-configurable parameter or may be a "hard
coded" constant
▪ A combination of the above
- For the selected structure, voxels having dose values close to D_selected may be identified as dose modification voxels
∘ Voxels may be identified as dose modification voxels if the voxels have values falling
within D_selected +/- Δ
▪ Δ may be a fixed value (which may be operator-selectable) or a fraction of D_selected (which may be an operator-selectable fraction)
∘ If no voxels are identified to be dose modification voxels then Δ may be expanded
and voxels may be re-identified
▪ This process repeats until at least one voxel is identified as a dose modification
voxel
∘ If a large number of voxels (e.g. all voxels for that structure or a number of voxels
greater than a threshold number or a threshold percentage of the voxels in a structure)
are identified to be dose modification voxels, then Δ may be reduced and voxels may
be re-identified
∘ In some examples, all voxels inside the structure for which the DVH is changed may
be identified as dose modification voxels
- Optional: identification of secondary dose modification voxels in marginal region
around primary dose modification voxels and assignment of suitable magnitude of dose
modification to secondary dose modification voxels
- Dose modification is performed for voxels identified as dose modification voxels (see
Dose Modification below)
- Once Dose Modification is performed, DVH is re-calculated based on new estimated dose
- Check whether any dose restrictions are violated (see Dose Manipulation Under Dose
Restrictions)
- Display is updated with new DVH
Dose Modification
[0110]
- For each voxel identified to be a dose modification voxel (Vox_identified) and for each beam:
∘ Determine the location(s) where ray line(s) passing through Vox_identified intersect with the beamlet matrix
∘ Modify intensity values of beamlet(s) at intersection point(s) (dose-change beamlets)
using a default quantity
▪ The default quantity may be correlated with (e.g. proportional to) the magnitude
of the desired dose change and may also depend on the number of dose-change beamlets
who ray lines intersect Vox_identified
∘ Optional: identify and determine intensity value adjustment for secondary dose-change
beamlets in a marginal region around primary dose-change beamlets
- Modified beamlets are evaluated for possible intensity restriction violations. By
way of non-limiting example:
∘ Intensity values that are less than a minimum threshold
▪ Corrected by increasing violating beamlet intensities to minimum threshold
∘ Intensity values greater than a maximum threshold
▪ Corrected by decreasing violating beamlet intensities to maximum threshold
∘ Variations in intensity over the two dimensional extent of the beamlet matrix may
also be limited (e.g. a maximum variation of a fixed percentage from one beamlet to
an adjacent beamlet)
▪ Corrected by increasing or decreasing the intensities of modified beamlets so that
variation restriction is no longer violated
- Dose is estimated at voxels for all beams using Rapid Estimation of Achievable Dose
Distribution
∘ Dose may be calculated for only a subset of voxels. For example:
▪ Voxels of the selected structure only
▪ Voxels of all structures
▪ Voxels of all structures plus currently displayed cross sectional plane
- Estimated dose for each dose modification voxel (Vox_identified) is compared to dose requested by operator:
∘ Determine difference between the mean estimated dose and mean operator-requested
dose for all dose modification voxels Vox_identified
∘ Rescale magnitude of intensity modification for all dose-change beamlets is rescaled
so that mean dose for each dose modification voxel (Vox_identified) is equal to mean operator-requested dose requested by operator unless the rescaling
of the dose-change beamlets results in a violation of the beamlet intensity restrictions
Dose Manipulation Under Dose Restrictions
[0111]
- Restrictions on the shape of DVH curves (or other restrictions on the dose distribution
itself or on other dose quality metrics) may be designated by the operator or may
be otherwise incorporated into the procedure (e.g. as global parameters of the system/method).
For example:
∘ No voxels within a structure can be below a specified dose threshold
∘ No voxels within a structure can exceed a specified dose threshold
∘ A specified percentage of volume of a structure cannot exceed a specified dose threshold
∘ A specified percentage of volume of a structure cannot exceed a specified dose threshold
- When a dose manipulation is performed (through DVH or otherwise) the dose distribution
is evaluated for potential dose restriction violations
- Violations are detected by comparing the estimated dose distribution and/or dose quality
metrics (e.g. DVH) with the dose restrictions
∘ Voxels contributing to the violation are identified
∘ The magnitude of dose change required for each violating voxel to satisfy the restriction
is calculated
- Perform Dose Modification (see above) using the violating voxels and voxel dose change
requirements as inputs
- Repeat as necessary until there are no further dose restriction violations
Example Application
[0112] The following non-limiting example is meant to provide further understanding of how
various aspects and features of the disclosure could be used in practice. Figure 19A
is a three-dimensional rendering of a target structure 281 and healthy tissue structure
282 and Figure 19B is a cross-sectional view bisecting (and showing outlines of) target
structure 281 and healthy tissue structure 282 along the plane 283. Figure 19B shows
the x and y axes scales in millimeters.
[0113] The beam configuration used in this particular example comprises a 360° degree rotation
of the radiation source with respect to the subject. Such a beam configuration could
be implemented using the radiation delivery apparatus of Figures 1 or 2, for example.
In this example, it is assume that the radiation source moves in the x-y plane of
Figures 19A, 19B. The initial estimated dose (e.g. block 46) is generated using 180
beams distributed along the path of the radiation source. The initial estimate of
the dose distribution is determined using initial intensity distributions (e.g. block
44) wherein the initial beamlet intensities are proportional to the target prescription
dose for ray lines that project through target structure 281 and zero otherwise. In
this example, the target prescription dose is 60 Gy. The initial dose is estimated
using the convolution based rapid dose estimation technique described above, with
the dose estimate kernel according to equation (1) and using the inverse Fourier transform
method of equation (2). At each voxel, the total estimated dose is the sum of dose
contributions from each beam to that voxel. Figure 20A shows a representation of a
cross-section 291 of the initial estimated dose distribution for the Figure 19A target
tissue structure 281 and healthy tissue structure 282. Figure 20B shows the DVH for
the full three-dimensional target structure 292 and healthy tissue structure 293.
The Figure 20A dose distribution 291 is depicted as a grayscale pixel map, with lighter
pixels indicating higher dose. The initial dose estimate 291 is fairly uniform in
the volume of target structure 281 with the healthy tissue structure 282 receiving
substantial dose in the region adjacent to target structure 281.
[0114] Once the initial dose estimate is established, manipulation of the estimated dose
distribution may be permitted (e.g. by an operator). The operator may desire to establish
adequate volumetric coverage of target structure 281 with the prescription dose. This
may be accomplished by increasing the minimum dose to target tissue structure 281
through manipulation of the DVH. Figures 21A, 21B and 21C show an example of increasing
the minimum dose to target structure 281 through DVH manipulation. Figure 21A shows
a magnified portion 301 of the Figure 19B DVH plot 292. The operator may select the
DVH curve near a low dose point 303 on the curve 301 (e.g. using the computer mouse
or similar computer pointing device), depress the mouse button, and then translate
curve 301 to the left (towards lower dose) or to the right (towards higher dose).
In this example, the operator increases the dose by moving the selected point 303
on DVH curve 301 to the right as shown at 304. When the operator moves the DVH curve
in this manner, the dose modification voxels (and corresponding dose modification
magnitudes) are determined and a dose modification is applied to those dose modification
voxels. In this example, the Figure 21A manipulation corresponds to three dose modificiation
voxels. When the dose modification is applied to these three dose modification, voxels
it is determined that no intensity limits will be violated. Figure 21B shows examples
of three dose modification distributions 305 resulting from the Figure 21A dose modification
request communicated by the operator (i.e. from the three dose modification voxels).
In the Figure 21B representations, lighter grayscale indicates higher dose. It may
be observed that each modification 305 is well localized to specific areas of the
dose distribution. Figure 21C shows an updated dose distribution and correspondingly
updated DVHs for the target structure 307 and healthy tissue structure 308. The updated
target structure DVH 307 of Figure 21C shows that the minimum dose to the target structure
281 is higher than it was preceding the modification (see DVH 292 of Figure 20B).
Also, it may be observed that healthy tissue DVH 308 has not changed substantially
(compare to healthy tissue structure DVH 293 of Figure 20B), which is a result of
the changes being localized to target structure 281 during dose modification.
[0115] Now that the operator has established an acceptable dose distribution for target
structure 281, it may be desirable to place restrictions on DVH manipulation so that
the acceptable dose distribution is maintained during further dose modification. For
example, such a restriction could be that 100% of the target structure 281 should
receive 57 Gy and 0% of the target structure 281 should receive more than 63 Gy. The
operator may then proceed with modifying the DVH 315 corresponding to healthy tissue
structure 282 by selecting a point on DVH 315 and dragging the mouse or similar computer
pointing device to the left 317 (for dose reduction). Once the dose modification voxels
corresponding to this desired DVH modification have been determined, a dose modification
is applied to each dose modification voxel. In this case restrictions on the intensity
changes necessitate the use of a smaller number of beamlets. Figure 22B shows corresponding
dose modification distributions 318 which may be different from those of Figure 21B
due to the smaller number of beamlets used. Note that the dose modification distributions
318 are well localized inside the healthy tissue structure 282. Figure 22C shows the
updated DVH 319. Figure 22C shows that the dose to healthy tissue structure 282 has
been reduced as desired.
[0116] With the healthy tissue dose reduction illustrated in Figure 22C, the restriction
321 placed on the minimum dose for target tissue structure 281 is violated (see Figure
22C). A corrective dose modification may be automatically applied to bring the target
structure DVH back within the constraint 321. The dose modification voxels where the
dose distribution causes the violation of the DVH constraint 321 are determined and
dose modifications are applied at those dose modification voxels. Figure 23A shows
two example dose modification distributions 323 used to correct the violation of constraint
321 on the target dose distribution. Figure 23A shows that in this case the dose modification
is applied at the edge of target structure 281 close to healthy tissue structure 282
where the Figure 22 dose decrease was applied. Again the dose modification distributions
are well localized so that areas surrounding the dose modification voxels are not
substantially impacted. Figure 23B shows target structure DVH 324 after application
of the Figure 23A dose modification 323 and how target structure DVH 324 is now within
the restrictions (321 and 314). The healthy tissue structure DVH 325 has changed only
slightly from that of Figure 22C due to the dose changes being localized primarily
within the target structure 281. Figure 23C shows a representation of the resultant
cross-sectional dose distribution 326 which shows a reduction in dose to the healthy
tissue structure compared to that of Figure 21C, as was originally desired by the
operator.
[0117] Further dose modifications may be imposed by the operator by repeating in a similar
fashion the process described above. For example, further reductions in the healthy
tissue structure dose may be applied under the constraints 321 and 314 (Figure 23B).
Eventually an equilibrium will be reached where further reduction in healthy tissue
structure dose cannot be achieved without violating the restrictions on the target
dose. Figure 24A shows the DVH for the target 331 and healthy tissue structure 332
at the end result of a series of dose modifications by the operator. Figure 24B shows
the corresponding cross-sectional estimated dose distribution 333 corresponding to
these modifications. The DVH 332 and estimated cross-sectional dose distribution 333
for the healthy tissue structure 282 have been reduced considerably as compare to
the initial DVH 293 (Figure 20B) and dose distribution 291 (Figure 20A). A high uniform
dose is maintained in the target as seen in the final target DVH 331 (Figure 24A)
and cross-sectional dose distribution 333 (Figure 24B).
[0118] After the dose manipulation is complete the radiation delivery parameters for the
beam configuration may be determined. The radiation delivery parameters may then be
transferred to the control system and computer of the radiation delivery apparatus.
The radiation may then be delivered to the subject, thereby delivering a dose distribution
in the subject substantially similar to that derived from the dose manipulation.
[0119] Embodiments of the present invention include various operations, which are described
herein. These operations may be performed by hardware components, software, firmware,
or a combination thereof.
[0120] Certain embodiments may be implemented as a computer program product that may include
instructions stored on a machine-readable medium. These instructions may be used to
program a general-purpose or special-purpose processor to perform the described operations.
A machine-readable medium includes any mechanism for storing information in a form
(e.g., software, processing application) readable by a machine (e.g., a computer).
The machine-readable medium may include, but is not limited to, magnetic storage medium
(e.g., floppy diskette); optical storage medium (e.g., CD-ROM); magneto-optical storage
medium; read-only memory (ROM); random-access memory (RAM); erasable programmable
memory (e.g., EPROM and EEPROM); flash memory; or another type of medium suitable
for storing electronic instructions.
[0121] Additionally, some embodiments may be practiced in distributed computing environments
where the machine-readable medium is stored on and/or executed by more than one computer
system. In addition, the information transferred between computer systems may either
be pulled or pushed across the communication medium connecting the computer systems.
[0122] Computer processing components used in implementation of various embodiments of the
invention include one or more general-purpose processing devices such as a microprocessor
or central processing unit, a controller, graphical processing unit (GPU), cell computer,
or the like. Alternatively, such digital processing components may include one or
more special-purpose processing devices such as a digital signal processor (DSP),
an application specific integrated circuit (ASIC), a field programmable gate array
(FPGA), or the like. In particular embodiments, for example, the digital processing
device may be a network processor having multiple processors including a core unit
and multiple microengines. Additionally, the digital processing device may include
any combination of general-purpose processing device(s) and special-purpose processing
device(s).
[0123] Although the operations of the method(s) herein are shown and described in a particular
order, the order of the operations of each method may be altered so that certain operations
may be performed in an inverse order or so that certain operation may be performed,
at least in part, concurrently with other operations. In another embodiment, instructions
or sub-operations of distinct operations may be in an intermittent and/or alternating
manner.
[0124] Where a component (e.g. a software module, processor, assembly, device, circuit,
etc.) is referred to above, unless otherwise indicated, reference to that component
(including a reference to a "means") should be interpreted as including as equivalents
of that component any component which performs the function of the described component
(i.e. that is functionally equivalent), including components which are not structurally
equivalent to the disclosed structure which performs the function in the illustrated
exemplary examples of the invention.
[0125] As will be apparent to those skilled in the art in light of the foregoing disclosure,
many alterations and modifications are possible in the practice of this invention
without departing from the scope thereof. For example:
- The above-described examples describe several features which may be characteristics
of each beam 159. By way of non-limiting example, the procedures of block 72-78 for
each loop of method 70 (Figure 13) are performed for each beam 159 and Figure 12 shows
a number of beams 159. In some examples, radiation delivery may involve continuous
movement of the radiation source with respect to the subject over a given trajectory.
In such examples, the trajectory of movement of the radiation source with respect
to the subject may be described by a number of discrete sample beams from along trajectory.
Such sample beams may be sampled sufficiently closely to one another such that the
sum of the dose contributions of the sample beams is representative of the sum of
the dose contributions from the continuously moving radiation source. Such sample
beams may be used in a manner similar to that described herein for conventional discrete
beams and unless specifically described as being discrete, any references to beam(s)
herein should be understood to include sample beams.
- Referring to Figure 4A, it will be appreciated by those skilled in the art that the
number of beam 159 is not required to be the same in blocks 142 and 143. For example,
it can be computationally more efficient to have a relatively small number of beams
159 when permitting manipulation of the estimated dose in block 142. Then, in block
143, a larger number of beams 159 can be used to accurately determine the radiation
delivery parameters. In some examples, the number and/or location of beams 159 may
be one of the variables that is permitted to change in a block 143 optimization process.
- The discussion presented above suggests that there is a one to one relationship between
beamlets 164 and ray lines 163 - i.e. each beamlet 164 has a corresponding ray line
163. Such ray lines 163 may pass through the center of their corresponding beamlets
164. Using such ray lines 163, a relationship can be established between particular
beamlets and particular voxels - e.g. dose -change beamlets can be identified as those
beamlets that have corresponding ray lines that project through a dose modification
voxel (see discussion of block 221 above). It will be appreciated, however, that beamlets
actually have a two-dimensional shape. In some examples, it may be desirable to model
this two dimensional beamlet shape by accounting for the impact of a beamlet on voxels
other than merely the particular voxels through which its corresponding ray line passes.
For example, in some rapid dose estimation examples (e.g. in some examples of method
70 of Figure 13) the ray lines associated with particular convolved intensity values
may impinge on particular voxels at locations away from the centers of the particular
voxels and the dose contributions from those convolved intensity values may be fractionally
divided as between the particular voxels and one or more of their neighboring voxels.
The fractional division of dose contributions between the particular voxels and neighboring
voxels may depend on the location where the ray line impinges on the particular voxels.
For example, most of the fractional dose contribution may be added to a particular
voxel intersected by a ray line and the remaining dose contribution may be divided
between the neighboring voxels, with neighboring voxels relatively close to the location
of intersection receiving a greater dose contribution than neighboring voxels that
are relatively far from the location of intersection.
- In some rapid dose estimation examples (e.g. in some examples of method 70 of Figure
13), it may be the case that multiple ray lines 163 from a single beam of convolved
intensities impinge on a single voxel. In such cases, the dose contribution added
to the voxel may be some form of average or interpolation of the dose contributions
that would be predicted by each of the individual convolved intensity values.
- The discussion presented above suggests that there is a one to one relationship between
beamlets 164 and ray lines 163 - i.e. each beamlet 164 has a corresponding ray line
163 which may pass through its center. However, in other examples, for each beam,
ray lines could additionally or alternatively have a one to one relationship with
voxels - i.e. each voxel could have a corresponding ray line that extends from the
center of the voxel toward the radiation source location for that particular beam.
In such examples, the block 221 (Figure 4B) process for identification of dose-change
beamlets could involve tracing rays from the block 220 dose modification voxels onto
the two-dimensional grids 162 of beamlets 164 - i.e. dose-change beamlets could be
identified as those through which a ray line projects. In some examples, it may be
desirable to model the three-dimensional voxel shape by accounting for the impact
of a voxel on beamlets other than merely the particular beamlets through which its
corresponding ray line passes. For example, in some examples of method 18 (Figure
4B), the ray lines associated with particular dose modification voxels may impinge
on particular beamlets at locations away from the centers of the particular beamlets
and the block 223 intensity value adjustments may be fractionally divided as between
the particular beamlets and one or more of their neighboring beamlets. The fractional
division of intensity adjustments between the particular beamlets and neighboring
beamlets may depend on the location where the ray line impinges on the particular
beamlets. For example, most of the fractional intensity value adjustment may be made
to a particular beamlet intersected by a ray line and the remaining intensity value
adjustment may be divided between the neighboring beamlets, with neighboring beamlets
relatively close to the location of intersection receiving a greater fraction of the
intensity value adjustment than neighboring beamlets that are relatively far from
the location of intersection.
- In some examples, ray lines for each beam may have a one to one association with voxels
and ray lines may extend from the center of voxels to the radiation source location
for that particular beam. In such examples, it may be the case (e.g. in block 221
of method 18 (Figure 4B)) that multiple ray lines from multiple voxels impinge on
a single beamlet. In such cases, the block 223 intensity value adjustment to the beamlet
may be some form of average or interpolation of the intensity value adjustments that
would be applied by each of the individual voxels.
- One technique for rapid estimation of achievable dose described above involves convolving
a two-dimensional intensity distribution i(x,y) and the two-dimensional dose estimate kernel k(x,y) and then projecting the convolved intensity values of the resultant two-dimensional
convolved intensity distribution f(x,y) along corresponding ray lines by adding the convolved intensity values as dose contributions
to voxels intersected by the ray lines. In some examples, it may be desirable to compute
convolved intensity distributions f1(x,y), f2(x,y), ... for a number of different dose estimate kernels k1(x,y), k2(x,y), ..., where the different dose estimate kernels can be used to model different tissue
densities and their corresponding different scattering patterns. Then, when estimating
achievable dose for a particular voxel (e.g. when adding convolved intensity values
as dose contributions to the particular voxel), the tissue density of the particular
voxel p can be used to select convolved intensity values from a convolved intensity distribution
fp(x,y) computed using a dose estimate kernel kp(x,y) which corresponds to the tissue density.
[0126] In the foregoing description, the invention has been described with reference to
specific embodiments and examples thereof. It will, however, be evident that various
modifications and changes may be made thereto without departing from the scope of
the invention, which is defined by the appended claims. The description and drawings
are, accordingly, to be regarded in an illustrative sense rather than a restrictive
sense. While a number of exemplary aspects and examples have been discussed above,
those of skill in the art will recognize certain modifications, permutations, additions
and sub-combinations thereof. It is therefore intended that the following appended
claims and claims hereafter introduced are interpreted to include all such modifications,
permutations, additions and sub-combinations as are within their scope.
1. Computerimplementiertes Verfahren (18), um Manipulation einer erzielbaren Schätzung
der Dosisverteilung zu erlauben, die durch eine Strahlungsabgabevorrichtung für die
vorgeschlagene Behandlung einer Patientenperson lieferbar ist, wobei die erzielbare
Abschätzung der Dosisverteilung über einen dreidimensionalen Bereich von Voxeln mit
einem Dosiswert für jedes Voxel definiert wird, wobei das Verfahren umfasst:
(a) Ermitteln (220)eines Voxels zur Dosisänderung , für das es gewünscht ist, den
Dosiswert zu ändern und Ermitteln einer entsprechenden Größenordnung der gewünschten
Dosisänderung;
(b) für jeden einer Vielzahl von Strahlen:
(i) Charakterisieren des Strahls als eine zweidimensionale Anordnung von Beamlets
(kleinen Strahlen), wobei jedes Beamlet mit einem entsprechenden Intensitätswert verbunden
ist und wobei eine Strahllinie, die eine Projektion des Beamlets in den Raum darstellt;
und
(ii) Identifizieren (221) eines oder mehrerer Beamlets zur Dosisveränderung, die zugehörige
Strahllinien aufweisen, die das Voxel für Dosisänderung schneiden;
(c) Ändern (223) der Intensitätswerte der identifizierten Beamlets für Dosisveränderung
auf der Größenordnung der gewünschten Dosisänderung beruhend, wobei die Änderung des
Intensitätswertes der identifizierten Beamlets für Dosisveränderung, für jedes der
identifizierten Beamlets für Dosisveränderung, das Ändern des Intensitätswertes des
identifizierten Beamlets für Dosisveränderung gemäß einer Funktion umfasst, die proportional
zur Größenordnung der gewünschten Dosisänderung und inverse proportional zur Zahl
der identifizierten Beamlets für Dosisveränderung ist; und
(d) Aktualisierung (224) der erzielbaren Schätzung für Dosisverteilung auf Basis der
geänderten Intensitätswerte der identifizierten Beamlets für Dosisveränderung.
2. Computerimplementiertes Verfahren nach Anspruch 1, wobei das Ermitteln des Voxels
zur Dosisänderung umfasst:
Empfangen, als Eingabe, einer angeforderten Änderung einer Dosisqualitätsmetrik; und
Verwenden der angeforderten Änderung der Dosisqualitätsmetrik, um das Voxel für Dosisänderung
zu ermitteln.
3. Computerimplementiertes Verfahren nach Anspruch 2, wobei die Dosisqualitätsmetrik
eine Dosisvolumen-Histogrammkurve (DVH-Kurve) umfasst, und wobei die Verwendung der
angeforderten Änderung der Dosisqualitätsmetrik zum Ermitteln des Voxels für Dosisänderung
umfasst:
Identifizieren einer Stelle der angeforderten Änderung auf der DVH-Kurve, um einen
Dosiswert D_selektiert zu haben; und
Identifizieren eines Voxels in der erzielbaren Abschätzung der Dosisverteilung, die
einen Dosiswert in einem Bereich D_selektiert ± Δ aufweist, der das Voxel für die Dosisänderung sein soll, wobei Δ ein Bereichsparameter
ist.
4. Computerimplementiertes Verfahren nach Anspruch 1, wobei das Ermitteln der Größenordnung
der gewünschten Dosisänderung das Ermitteln der Größenordnung umfasst eine oder mehrere
von folgenden zu sein: Ein Bruchteil f(0≤f≤1) des Dosiswertes für das Voxel zur Dosisänderung;
ein Bruchteil f(0≤f≤1) einer maximalen Dosisbeschränkung für eine Gewebestruktur an
der Stelle des Voxels zur Dosisänderung; ein Bruchteil f(0≤f≤1) einer verschreibungspflichtigen
Dosismenge für eine Zielgewebestruktur an der Stelle des Voxels für Dosisänderung;
einer Menge proportional zur Bewegung einer Computer-Anzeigevorrichtung; einer Menge,
die von einem Bediener eingegeben wurde; einer vom Bediener konfigurierbaren festen
Menge; einer festen Menge, die ein Parameter eines Systems ist, auf dem das Verfahren
durchgeführt wird.
5. Computerimplementiertes Verfahren nach einem der Ansprüche 1 bis 4, das für jeden
der Vielzahl von Strahlen und für jedes des einen oder der mehreren Beamlets für Dosisveränderung,
das Ermitteln eines oder mehrerer sekundärer Beamlets für Dosisveränderung in einem
marginalen Bereich nahe dem Beamlet für Dosisveränderung und das Behandeln der sekundären
Beamlets für Dosisänderung als Beamlets für Dosisveränderung bei Durchführung der
Schritte (c) und (d) umfasst.
6. Computerimplementiertes Verfahren nach einem der Ansprüche 1 bis 5, wobei die Aktualisierung
der erzielbaren Schätzung der Dosisverteilung für jeden der Vielzahl von Strahlen,
die ein oder mehrere Beamlets für Dosisveränderung umfassen, umfasst:
Falten des Intensitätswertes der zweidimensionalen Anordnung von Beamlets mit einem
zweidimensionalen Dosisschätzungs-Kernel, um eine zweidimensionale gefaltete Intensitätsverteilung
zu erhalten, wobei die zweidimensionale gefaltete Intensitätsverteilung einen gefalteten
Intensitätswert für jedes Beamlet umfasst; und
für jedes Beamlet in der zweidimensionalen Anordnung von Beamlets: Identifizieren
von Voxels, die von der Strahllinie geschnitten werden, die mit dem Beamlet verbunden
ist; und Hinzufügen eines Dosisbeitrags zu den geschnittenen Voxels, wobei der hinzugefügte
Dosisbeitrag auf dem gefalteten Intensitätswert des Beamlets basiert.
7. Computerimplementiertes Verfahren nach Anspruch 6, wobei der Dosisschätzungs-Kernel
eine zweidimensionale Punktausbreitungsfunktion umfasst, und wobei das Falten der
Intensitätswerte der zweidimensionalen Anordnung von Beamlets mit dem zweidimensionalen
Dosisschätzungs-Kernel umfasst: Multiplizieren einer Fourier-Transformation der Intensitätswerte
der zweidimensionalen Anordnung von Beamlets mit einer Fourier-Transformation des
zweidimensionalen Dosisschätzungs-Kernels, um Fourier-Multiplikationsergebnis zu erhalten;
und Ermitteln einer inversen Fourier-Transformation des Fourier-Multiplikationsergebnisses,
um die zweidimensionale gefaltete Intensitätsverteilung zu erhalten.
8. Computerimplementiertes Verfahren nach einem der Ansprüche 1 bis 7, wobei die Aktualisierung
der erzielbaren Abschätzung der Dosisverteilung umfasst, dass die aktualisierte erzielbare
Dosisverteilung einer oder mehrerer Begrenzungen unterzogen wird.
9. Computerimplementiertes Verfahren nach Anspruch 8, wobei, wenn die aktualisierte erzielbare
Dosisverteilung, die einer oder mehrerer Begrenzungen unterzogen wird, schlussfolgert,
dass eine Verletzung irgendeiner Dosisbeschränkung vorliegt, dann umfasst das Verfahren:
Angehen der Verletzung der Dosisbeschränkung, wobei das Angehen der Verletzung der
Dosisbeschränkung umfasst: Ermitteln eines oder mehrerer die Begrenzung verletzender
Voxel, die bewirken, dass die aktualisierte Dosisverteilung die Dosisbeschränkung
und eine oder mehrere entsprechende gewünschte Größenordnungen von Dosisveränderung
verletzt; Wiederholen der Schritte (b), (c) und (d) mit dem einen oder mehreren Voxel,
welche die Beschränkung verletzen und der einen oder mehreren entsprechend gewünschten
Größenordnungen für Dosisveränderung anstelle des Voxels für Dosisänderung und der
entsprechenden Größenordnung der gewünschten Dosisänderung; erneutes Unterziehen der
aktualisierten erzielbaren Dosisverteilung der einen oder mehreren Dosisbeschränkungen;
und
Wiederholen des Angehens der Verletzung der Dosisbeschränkung, bis das Unterziehen
der aktualisierten erzielbaren Dosisverteilung einer oder mehrerer Dosisbeschränkungen
schlussfolgert, dass keine Verletzung irgendeiner Dosisbeschränkung vorliegt.
10. Computerimplementiertes Verfahren nach einem der Ansprüche 1 bis 9, das, nach der
Aktualisierung der erzielbaren Schätzung der Dosisverteilung, das Ermitteln von einer
oder mehreren aktualisierten Dosisqualitätsmetriken umfasst, welche die aktualisierte
erzielbare Schätzung der Dosisverteilung verwenden.
11. Computerimplementiertes Verfahren nach Anspruch 10, wobei das Ermitteln einer oder
mehrerer aktualisierter Dosisqualitätsmetriken umfasst, dass zumindest eine aktualisierte
Dosisqualitätsmetrik einer oder mehreren Beschränkungen für Dosisqualitätsmetrik unterzogen
wird.
12. Computerimplementiertes Verfahren nach Anspruch 11, wobei, wenn die zumindest eine
aktualisierte Dosisqualitätsmetrik, die einer oder mehreren Beschränkungen für Dosisqualitätsmetrik
unterzogen wird, schlussfolgert, dass eine Verletzung irgendeiner Beschränkung für
Dosisqualitätsmetrik vorliegt, dann umfasst das Verfahren:
Angehen der Verletzung der Beschränkung für Dosisqualitätsmetrik, wobei das Angehen
der Verletzung der Beschränkung für Dosisqualitätsmetrik umfasst: Ermitteln eines
oder mehrerer die Beschränkung verletzender Voxel, die bewirken, dass die zumindest
eine aktualisierte Dosisqualitätsmetrik die Beschränkung für Dosisqualitätsmetrik
und eine oder mehrere entsprechende gewünschte Größenordnungen für Dosisveränderung
verletzt; Wiederholen der Schritte (b), (c) und (d) mit dem einen oder mehreren Voxel,
welche die Beschränkung verletzen und der einen oder mehreren entsprechend gewünschten
Größenordnungen für Dosisveränderung anstelle des Voxels für Dosisänderung und der
entsprechenden Größenordnung der gewünschten Dosisänderung; erneutes Unterziehen der
zumindest einen aktualisierten Dosisqualitätsmetrik der einen oder mehreren Beschränkungen
für Dosisqualitätsmetrik; und
Wiederholen des Angehens der Verletzung der Beschränkung für Dosisqualitätsmetrik,
bis das Unterziehen, der zumindest einen aktualisierten Dosisqualitätsmetrik einer
oder mehrerer Beschränkungen für Dosisqualitätsmetrik schlussfolgert, dass keine Verletzung
irgendeiner Beschränkung für Dosisqualitätsmetrik vorliegt.
13. Computerimplementiertes Verfahren nach einem der Ansprüche 1 bis 12, das, nach dem
Aktualisieren der erzielbaren Schätzung für Dosisverteilung, das Ermitteln eines oder
mehrerer Strahlungsabgabeparameter auf Basis der aktualisierten erzielbaren Dosisabschätzung,
des einen oder den mehreren Strahlungsabgabeparametern umfasst, die zur Verwendung
durch die Strahlungsabgabevorrichtung zur Behandlung der Patientenperson geeignet
sind.
14. Datenverarbeitungssystem zum Erlauben von Manipulation einer erzielbaren Schätzung
für Dosisverteilung, die durch eine Strahlungsabgabevorrichtung für die vorgeschlagene
Behandlung einer Patientenperson lieferbar ist, wobei die erzielbare Schätzung der
Dosisverteilung über einen dreidimensionalen Bereich von Voxeln mit einem Dosiswert
für jedes Voxel definiert wird, wobei das System einen Prozessor umfasst, der konfiguriert
ist, Folgendes zu tun:
Ermitteln (220) eines Voxels zur Dosisänderung, für das es gewünscht ist, den Dosiswert
zu ändern und einer entsprechenden Größenordnung der gewünschten Dosisänderung; für
jeden einer Vielzahl von Strahlen:
Charakterisieren des Strahls als eine zweidimensionale Anordnung von Beamlets, wobei
jedes Beamlet mit einem entsprechenden Intensitätswert verbunden ist und eine Stahllinie
eine Projektion des Beamlets in den Raum darstellt; und Identifizieren (221) eines
oder mehrerer Beamlets für Dosisveränderung, die zugehörige Strahllinien aufweisen,
die das Voxel für Dosisänderung schneiden;
Ändern (223) der Intensitätswerte der identifizierten Beamlets für Dosisveränderung
auf der Größenordnung der gewünschten Dosisänderung beruhend, wobei die Steuereinheit
konfiguriert ist, den Intensitätswert der identifizierten Beamlets für Dosisveränderung
zu ändern, indem, für jedes der identifizierten Beamlets für Dosisveränderung, der
Intensitätswertes des identifizierten Beamlets für Dosisveränderung gemäß einer Funktion
geändert wird, die proportional zur Größenordnung der gewünschten Dosisänderung und
inverse proportional zur Zahl der identifizierten Beamlets für Dosisveränderung ist;
und
Aktualisieren (225) der erzielbaren Schätzung für Dosisverteilung auf Basis der geänderten
Intensitätswerte der identifizierten Beamlets für Dosisveränderung.
15. Computerprogrammprodukt, das Befehle trägt, die in einem nicht flüchtigen computerlesbaren
Medium verkörpert sind, wobei die Befehle, wenn durch einen geeigneten Prozessor ausgeführt,
bewirken, dass der Prozessor das Verfahren nach irgendeinem der Ansprüche 1 bis 13
durchführt.